i '■]\-r':'i't'}'5) has expressly stated that the anchor-teeth, four in number to each head, are to be considered not as secondary spines but as real rays, — a view which is not acceptable, since the transverse axial canals as plainly shown in his figure are very short and far from extend- ing even to the base of the teeth. On the other hand, an essentially different type of anchoring spicules — genuine pentactin anchors — has been discovered by F. E. SenuLZE as occasionally present, in addition to the barbed diactin type, in the root-tuft of certain Euplectella species (E. aspergillum, simplex, asper). The unpaired ray is prolonged into the shaft which is perfectly smooth, while the short cruciately disposed rays at the inferior end are recurved and form the anchor- tee til. The latter are each traversed throughout by the axial canal. EUPLECTELLA. 51 Hexasters, — These are, according to my conception, minute liexactins wliicli are invariably characterized by the presence of a number of slender, radiall}^ disposed secondary appendages — the terminal rays — at the outer end of each ray. The axial canal (PI. IV, fig. 20 ; PL V, figs. 30-34) is confined to the latter, which is called the principal ray, and never extends into the terminal ray, as can be easily demonstrated by examining the spicule in a medium whose refractive index approaches that of the siliceous matter. Not unfrequently the principal possesses but a single terminal by reduction, and w^hen the two are in a straight line, as is often the case, the external appearance is exactly like that of a simple primary ray. However, the ex- clusive presence of the axial canal only at the base will at once reveal the composite nature of such an apparently simple ray. The most constant form of hexasters in Eupledella is the floricome, which I regard as a variety of discohexasters. The terminal disc, instead of being uniformly developed all around, possesses strong marginal prongs only on the side turned away from the axis of the perianth of the terminals, while on the opposite inner side the disc-edge remains smooth and obtusely rounded, being only indicated by a hump-like curvature of the surf\ice (PI. II, fig. 14, d ; &c.). A parallel case of the same modification is found in a new octasterophorous Kossellid, which will be described under the name of Rhabdocalyptus ungui- culatus ; in this, the discoctaster exhibits a similar hand-like development of the terminal discs. In their general shape the floricomes show no noteworthy variation in the different species of the genus. Therefore, in this respect, as also in that of the number of terminals in a perianth or of marginal prongs on the terminal plate, they are scarcely 52 I. IJIMA : HEXACTINEELIDA. I. of importance in the systematic of the genus, save that their size has in certain cases been found to be of use in the specific distinction. The place and manner of origin of the floricome as well as its subsequent history seem not to have been followed out with accuracy by previous writers. It arises among the external trabecule beneatli the dermal latticework. At a certain stage of its development, the terminals are short and exceedingly fine (PL II, fig. 10, PL IV, fig. 11 ; &c.). By the time the perianths have reached the definitive shape and size by the elongation and flaring out of the still slender terminals, the entire rosette is of the form which has been called by F. E. Schulze the sigmato- come (PL II, fig. 11 ; PL IV, fig. 12) and which he has apparently taken for a separate category of hexasters in E. regalis (19', p. 28). The distal portion of the terminals continues to thicken ; then, the rudiment of the terminal disc is formed (PL II, figs. 12, 14 ; PL IV, figs. 13, 14), which stage in the development of the floricome has already been recognized by F. E. Schulze in E. aspera and Didyaulua clegans ('95, pp. 29, 41). However, that writer seems not to have observed its much earlier stage in which the terminals are still quite short, and evidently on that account, it appeared to him that all the radial rays attained their full length, though much more slender at first than in the later state, immediately upon their origin {I.e., p. 41). This is at any rate not quite true with the terminals of the floricome. They do grow gradually in length during their development, an obser- vation which I have found perfectly corroborated in the development of the graphiocome also (PL V, figs. 32-34). The floricome, after the complete development of its parts, seems not to be destined to remain at the locus nascendi but to EUPLECTELLA. 53 be normally moved oft' towards tlie external surface, finally to take a position at the extreme outer end of tlie distal rays of the dermalia (PL V, fig. 36). Analogous to the rhabditi of Turbellarians, the floricomes originate in deep parts and shift themselves over to the most superficial situations to effectively discharge their function as defensive weapons. This point in their history seems to have been hitherto entirely overlooked. As is well known, it is usual to find a floricome to the tip of each dermal hilt-ray on depressed and therefore more protected parts of the external surface ; whereas on more exposed parts, as on the ledges, it is frequently missing or exceedingly rare. The graphiocome may fairly be said to be tolerably constant in Eupledella, notwithstanding it has not yet been discovered in certain species {E. regalis, ciicumer, suberea, jovis & crassistellaia). I tliink that at least some of these species may yet be discovered to be not totally wanting in the said rosette. Whenever present, it is found, like the floricome, exclusively in the external trabec- ular layer. A remarkable fact, which has not been noticed by previous writers, in connection with the rosette in question, is, that the sheaves of the fine needle-like terminals are exceedingly liable to break off" close to the discs at the outer end of the principal rays, after the rosette has attained its full size. I believe that this breaking oft" is in fact a normal process, by which the rhaphides — a name that has been given to the liber- ated terminals without the knowledge of their genetic connec- tion with the graphiocome — are put in a position to be moved oft', with one of their ends pointed outward, towards the external surface, probably by the same force that drives the floricome in the same direction. Finally they are found in the most peri- pheral positions on the wall either scattered or in grouj)S and 54 I. IJIMA : HEXACTINELLIDA. I. especially iu intimate association with the hilt-rays of the dermalia, with their outer ends at, or sticking out of, the external bound- ing surface (PI. V, fig. 36). In such a situation and arrange- ment, the rhaphides would serve in their own way as a power- ful defense against attack from without. The excessive abundance of broken off oxyhexaster-terminals in perfectly undisturbed specimens of certain Rosselid species has given me the impression that the phenomenon is not peculiar to the graphiocome alone. The oxyhexasters (PL II, fig. 15; &c.) have been found in all Eu])lectella species, except in E. simplex and E. nodosa. Unlike the other rosettes already referred to, these occur iu both the external and the internal trabecular layer. In these positions they should serve as a defense against pernicious intruders. While in E. oweni and curvisiellaia the oxyhexasters were quite plentiful, I have found them rather sparingly in E. marshalli and imperialis and particularly so in the outer trabecular layer, — a circumstance which may be correlated with the especial abundance of graphiocomes in the two last mentioned species. The entire size of oxyhexasters and the different develop- ment of their parts are of value, though not universally, in the distinction of species. A very noteworthy modification of the oxy- hexaster is the clasp-like or sigma-like fibula of E. jovls (also found in another Eupectellid, Holascus fibulatus), the true nature of which has been perceived by F. E. Schulze ('87, pp. 78, 88). AVe have here to do with a diactinose oxyhexaster, probably derived from such a form as I have called the hexactin-shaped or hexactinose oxyhexaster (Ijima '97, p. 45 ; ' Derivate-Oxy- hexaster of Schulze, '99) in which each principal has only a single terminal, in a manner analogous to that in which a rod- EUPLECTELLA. 55 like diactin is derived from a simple hexactin by the suppres- sion of the four crueiately disposed ra3^s. Of special interest is ^. suberea W. Thoms. in that it gives some clue to the genetic relation between the oxyhexaster and a form of discohexasters, the onychaster (PL X, figs. 12, 20, 21), which appears strikingly like the former but is distinguished by having exceedingly fine claw-like or branch-like appendages at the outer end of the terminals. It is known with certainty that certain individuals of that species possess true oxyhexasters in abundance (Schulze '99, p. 19 ; Topsent '92, p. 24). Now, Topsent {I.e.) made the interesting observation that in a specimen examined by him the oxyhexasters were entirely replaced by onychasters, and further that in another specimen there occurred neither the one nor the other in typical development but a form combining the characters of both in that one or more of the terminals bore each a single hook at the free end, while the rest terminated in simple points. The onychaster was evidently also seen by W. Thompson in one or the other of the specimens obtained by the ' Challenger,' for we see one represented in fig. 8, PI. V, of the Challenger Report, which plate was prepared by that eminent naturalist, although in the text of the report Schulze held the spicule of that figure to be of extrinsic origin. With respect to what might be called the onycho-oxyhexaster discovered by Topsent, that writer justly concludes that it represents a form intermediate between discohexasters and oxyhexasters. Cases suggestive of the same transition are also known in the genus Aphrocallistes [A. ramosus, hocagei ; Schulze '95, pp. 77, 80). Which of the two hexaster-forms, the onychaster or the oxyhexaster, is then the more primitive ? Taken alone, the onycho- oxyhexaster would look just as much like an oxyhexaster on the 56 T. IJIMA : HEXACTINELLIDA. I. verge of developing into a discohexaster as like a discohexaster just before losing the last vestige of the terminal armature. Never- theless, T believe that discohexasters in the broad sense are all to bo regarded as having been derived from oxyhexasters by complica- tion of parts. I say this not on purely a priori grounds alone but also from the nature and mode of the growth of hexaster- terminals in general. Concerning the oxyhexaster of Eiqjiectella in particular or of the onychaster of any genus, I have made none or but little direct observation as to the development of the terminals. However, in certain Rossellid species I have not infrequently met with unusually small oxyhexasters of typical shapes, in which the terminals were presumably still growing. How these grow in the floricome (a discohexaster) and the graphiocome (an oxyhexaster), has already been referred to in brief and will be described in greater details under E. marshalli. These observations seem to sufficiently warrant the induction that the hexaster-terminals are in all cases secondary appendages at the outer end of the principals, or primary rays which alone inclose the axial filament, — local formations which at their inci- pient stage are exactly comparable to the simple microtubercles or spines that so frequently beset the spicular surface. It appears then assuraable that the discohexaster-terminals originally ended in simple points as the oxyhexaster-terminals always do and that the terminal disc or whorl of claw-like processes is, so to speak, a tertiary structure added to the simple end of the terminals after these had grown to their full length, — a factual demonstration of which changes is found in the development of the floricome. Thus the ontogenetic and phylogenetic sequence of discohexasters to oxyhexasters in general seems to be plain, and the onycho- oxyhexaster probably represents a stage of the passage of the EUrLECTELLA. 57 latter into the former. And yet, in view of the indubitable cases of oxjhexasters reverting back again into the more primary liexactin-shape (liexactinose oxyhexaster), a somewhat analogous retrogression of discohexasters into oxyhexasters may be said to be not altogether impossible to imagine. However, if this really takes place at all, it must be of casual occurrence and would re- quire special circumstantial evidence in order to be recognized as such. The presence of the incipient forms of discohexasters, i.e., the on}? chaste rs, among isolated members of different families (Euplectellidse, Melittionidse) would probably require for its explanation an assumption of its independent origination at separate points in the phylogeny by convergent adaptation. In all thirteen species of Euplectella are to be considered as known at present. I will on the next page annex a key to all these species, which should bring out the main points of their structural differences and indicate to some extent the affinities existing among them. 58 I. IJIMA : HEXACTINELLIDA. I. Key to the Species of Euplectella. «• — Tlie parenchyma] principalla of tlie circular and longitudinal skeletal beams are staur- actins. a^. Spicules of the skeletal beams never in fusion. «'• Parietal ledges well developed; oscularia of miscellaneous forms E. marshalli Ij. (Sagami Sea; Suruga Gulf). 6*. Parietal ledges little or not at all developed ; oscularia mainly diactins. a^. Oxyhexaster with straight terminals, 50-70 [j- in diameter E. oweni Herkl. & Maksh. (NW. of Kyüshyü). b^. Oxyhexaster with terminal rays bent near the outer end, 75-100 pi in diameter. E. curvislellata Ij. (S. of Kyüshü). 6'. Spicules of the skeletal beams in fusion, at least in the basal region. c". Parietal ledges not developed ; all meshes of the skeletal framework with a parietal osculum each; oxyhexaster not present E. simplex F. E. ScH. (Sea of Bengal). d^. Parietal ledges or protuberances well developed ; a number of the meshes of the skeletal framework without parietal^oscula ; oxyhexaster present, c'. Ledges cut up into irregular knobs, flaps, &c.; oscularia chiefly hexactins and pentactins E. impevialis Ij. (Sagami Sea; Suruga Gulf). d^. Ledges with tolerably sharp continuous edges. a*. Ledges low ; without graphiocome...JE'. regcdis F. E. Sch. (Sea of Bengal). b*. Ledges prominent ; with grapliiocome...Ê'. a'^pergillum Ovf. (Philippines). ''. — Tile parenchymal principalia of the circular and longitudinal skeletal beams are all or at least partially oxypentactins, or oxyhexactins with one ray directed distally and radially. c'. The parenchymal principalia are oxyhexactins with a reduced proximal ray. Stauractins may occur in addition. c". Distal ray of parenchymal oxyhexactins armed with prongs and conspicuously projecting out of the external surface. Oxyhexaster with short slender principals and long terminals E. aspera F. E. Sch. (Indian Ocean). /-. Distal ray of parenchymal oxyhexactins smooth, not freely projecting (?) ; oxy- hexaster with moderately long and thick principals and short terminals E. cmssklellala F. E. Sch. (Mid-Pacific). d'. The parenchymal principalia are smooth oxypentactins with entirely suppressed proximal rays. However, oxyhexactins may occur in addition. g'^. Distal ray of parenchymal oxypentactins reaching to the external surface but not beyond; yet, with outwardly projecting bundles of small thin diactins; no oxyliexaster E. nodom F. E. Sen. (Bermudas). /('-. Distal ray of parenchymal oxypentactins freely projecting out of tlie external surface. e^. Oxyhexaster represented by clasp-like or sigma-like fibulre (oxydiaster) ; oscularia scepter-like monactins E. jovis O. Schm. (W. Indies). /'. Eitiier typical oxyhexaster or onychaster present; no fibula;. c*. Body tubular, scarcely bellied; spicules nowhere in fusion; oscularia rough rod-like diactins E. suherea W. Thoms. (Atlantic). d*. Body distinctly bellied; spicules in fusion in certain parts; oscularia unknown F. cucumer Ow. (Seychelles). E. TMPERIALIS. — GEN. CHARACTERS. 59 EUPLECTELLA IMPERIALIS Tj. Pis. I & II. 1 EiipledcUa oweni, Schulze, '87, p. 81. Euplectella imperialis, Ijima, '94, p. 365. Ill F. E. Schulze's 'Challenger' Report (p. 81) it stands recorded that there was found among the Japanese Hexactinellida collected by Döderleix — in addition to a specimen of Euplectella oweni — a completely macerated and much injured skeleton (320 mm. long) of Euplectella, in which the spicules seemed to be loose above but below were fused into a firm latticework. This was assumed as belonging to a very large and old individual of E. oweni, in which the usually unfused spicules had become soldered together. To my knowledge such a fusion of spicules never takes place in the species mentioned (see anon under E. oweni). It therefore seems to me likely that Schulze had before him the specimen referred to by Döderlein ('83 p. 105) as having been obtained by purchase at Enoshima ; and that, particularly in view of the above mentioned character of the skeleton, it belonged to the species which I am now going to describe under the designation of E. imperialis. A preliminary account of this species was given in 1894 in the Zoologischer Anzeiger. Since that period no less than fifty specimens have passed through my hands, including all sizes from one of only 30 mm. up to a giant of 825 mm. in length. They were mostly collected by Kuma. To give the exact localities where they were obtained : The majority came from Yodomi (both Naka-no-Yodomi and Maye- 60 I. IJIMA : HEXACTINELLIDA. I. no-Yodomi; 313-548 fms. [572-1002 ra.]) anclOkinose (Inside and Outside by the Bislia mon -line/'' the Iwado-line, &c.; 235-313 fms. [429-572 m.]). At Yodomi, Mr. Tsuchida and myself had the good fortune to capture some with our own hands. Several specimens also came from Homba and a few small individuals from Gokeba. Further in 1899 Kuma obtained for Mr. Owston a specimen off Tago (on the western coast of the Prov. of Izu) in Suruga Gulf near the 200 fathoras-line ; this specimen was identified by me as belonging to the present species. I should put the bathymetrical range of E. imperîalis as at present known at 200-548 fms. (365-1002 m.). It is evidently an inhabitant of deeper waters than E. ^narshalli, which does not occur at a greater depth than 160 fms. Besides, the nature of the bottom differs with the two species, as is attested by the matter interlocked in the basal tuft. While in the case of E. imperialis this consists almost purely of volcanic mud or sand of a gray color in the dried state, in the case of the other species the included matter is invariably shelly. In the fresh state tbe color of the sponge is a pale yellow, often appearing rather dirty, being soiled by the mire of the bottom. To the same cause is to be ascribed the gra^dsh color assumed by some specimens on drying, which otherwise should become perfectly colorless. Preserved in spirit the natural color is dissolved away. General Chaeactees of Nearly or Quite Full- Grown Specimens. E. imperialh shows many points of close agreement with E. aspergillum in regard to external form and structure, indicating *I luive not put clown tliis line on the chart of I'l. XIV. It lies between the lines of Iwado and Songenzuka: ïogeyama -o- the Village of Bishamon. E. IMPERIALIS. — GEN. CHARACTEES. 61 a near relationship between the two. As in that species, the tubular body usually exhibits a simple, horn-like curvature, more or less pronounced according to individuals (PI. I). However, nearly straight forms or those only slightly bent in S-like or irregular curves are by no means uncommon among the larger specimens. The body is approximately circular in cross-section. This is at any rate constantly the case with the basal portion of the body ; the upper portion may show certain irregularities in this respect. When quite full-grown the body is broadest at the upper end and gradually narrows below towards the bulbous basal tuft (PI. I, fig. 1). In less advanced stages of growth, however, the shape is that of a slightly bellied tube, the broadest part being situated at or near the middle (figs. 2 & 3) ; otherwise, the breadth remains nearly the same from the broadest portion upwards to the upper extremity. It is clear that after a certain period of life, the growth concerns the upper region only, the low^er portion admitting of little or no growth on account of the soldering together of the main skeletal ele- ments, and that the continued growth in girth at the upper end after the growth in length has ceased, finally converts the original bellied tube into the cornucopia-like shape broadest at the top. Thus, the specimen shown in fig. 2 w^ould have yet to grow broader in the upper region in order to attain the definitive shape, such as that of fig. 1. I may say that up to the stage when the body has grown to a length of about 300 mm., the shape of a bellied tube is invariably retained. Not unfrequently, the body considerably exceeds that length — in some cases reach- ing nearly 500 mm. — without deviating from the shape just re- ferred to ; while, on the other hand, others (e. g., specimen D 62 T. IJIMA : HEXACTINELLIDA. I. of the list appended below) may have already acquired the cornucopia-like shape when not much over a foot in height. The variation evidently stands in relation to that of maximum size, which different individuals are destined to attain. The following are dimensions of several specimens which I have measured : Total length, basal bulb inclu- sive. Height of body exposed above the sea-bottom. o i -; Diameter at middle of body exclusive of parietal pro- tuberances. Diameter just below the cuff. Eemarks on body-shape. A mm. 175 mm. 145 mm. 16 mm. 25 mm. 17 Bellied-tubiilar. C nr ved. B 230 190 18 25 19 jj » C 250 230 22 30 19 „ Nearly str aiglit. D 387 335 28 41 53 Connicüpia-like. Curved. E 1 390 3.57 25 54-51 44 Bellied-tiibiilar. Nearly str aight. F 465 420 24 41-45 42 >J 3> >7 5? G ; 478 425 35 57-65 64-69 Cornucopia-like. Curved. H\ 490 436 25 37-45 34 Tubular. Nearly straight. I 515 445 29 49-58 58-69 Cornucopia-like. Curved. J 570 500 32 55-56 75-80 JJ 5> K 573 518 30 55-58 51-91 J5 J? L 810 710 39 78-81 110 Slightly bent. The thickness of the body- wall, exclusive of the parietal ledges or protuberances, does not exceed 2-372 mm. The diameter at the extreme lower end of the skeletal tube, inclosed in the basal tuft, is only about one-fifth of that at the upper end. E. IMPERIALIS. — GEN. CHARACTERS. 63 The parietal leclgci^ or proluherances are very well developed and constitute a very characteristic feature of the species. They may in general be described as irregular ridges which are sub- ject to frequent interruptions in their course and which, far from presenting even surfeces and edge-lines, are cut up into numerous tubercular, knob-like or flap-like protuberances that lend a peculiarly corrugated or jagged appearance to the sponge. In this respect the present species presents a striking contrast to its nearest allies, E. aspergillum and E. regalis. At places the nappy ledges are seen running, numbers of them together, in an oblique direction, one way or the other, or in two intersecting oblique systems ; in other places they may show an altogether irregular disposition, often bending, branching or anastomosing in their course. They are certainly less con- spicuously developed in young than in old specimens. In the former there exists immediately below the cuft' a narrow zone in which the ledges are scarcely or not at all developed (PI. II, figs. 4, 6, 7) ; in the latter these may extend right up to, and join the base of, the cuö'on its uuderside (PI. I, fig. 1). Towards the lower extremity, the body is usually denuded to a greater or less extent of its peripheral loose tissues and with these the ledges also (figs. 1-3), thus exposing the bundles of basal spicules apposed to the skeletal latticework. So far as the prominences occur on the parietes, they are either tolerably uniformly devel- oped all over the body or may show greatest development in the middle region. In the largest individual before me (spec. L of the appended table) some of them measure as much as 14 ram. in height above the level of the parietal oscula. Certain flat lappets of the parietal ledge, particularly those with a sharper edge, are distinguished from the rest by having 64 T. IJIMA : HEXACTINELLIDA. I. a row of isolated bristle-like spicules standins; out along the free edge to a length of about 4 mm. In some specimens such fringed lappets are not at all uncommon, while in others they occur only occasionally and may even be entirely missing. Rounded protuberances never exhibit the prostal spicules. The parietal ledges form a rather steep wall to the valley- like, depressed spaces between them. These spaces are elongated or irregular in configuration, — generally meandering and inter- communicating, their shape depending upon the course taken by the inclosing ledges. Some of the more extensive, depressed areas may be said to have a comparatively flat surface. The parietal oscula, which do not exceed 2 mm. in diameter, open on the depressed areas. Their thin edges lie nearly on a level with the general surface of the latter. They are usually found several together in the same area at intervals of 2-10 mm. from one another. The distribution is on the whole irregular, though often a number of them in succession are found in a line, the direction of which depends ui:)on that of the long axis of the depression containing them. Exceptionally, isolated parietal oscula may open by means of a canal on the side or even on the summit of the external ledge. In specimens which are still actively growing at the upper end (figs. 3, 4), the openings are in that region arranged more or less regularly in transverse and longitudinal rows at short intervals. This regularity is however lost as the development of parietal ledges advances in that region. The extremely delicate dermal latticework is, on close examination, just visible to the' naked eye, except on the more elevated portions of the ledges, where the surface presents a rather close-grained texture. The apertures into the incurrent E. IMPERIALIS. — GEN. CHARACTERS. 65 canals, visible through the dermal layer, reach 1 inm. or slightly- more in diameter. The cuff (PI. II, fig. 8) is thin and of varying width owing to the irregularly undulating character of the edge-line. The width also varies with the size of the specimens. In the largest individual before me (spec. L of the list on p. 62) it is 11-20 mm. as measured on the upper surface ; 4-10 mm. in a specimen (G) 478 ram. high ; and only 2 mm. in a specimen (C) 250 mm. in height. The free edge may show at places the same interrupted fringe of marginalia as those found on certain lappets of the parietes. The fringe is however of inconstant occurrence. The surface of the cuff presents a close-grained appearance unless injured. The sieve-plate (PI. II, fig. 8) is usually arched like a watch- glass, but the convexity ma}^ in some individuals be more strongly pronounced than in others. The entire structure appears rather frail owing to the comparatively thin beams and large meshes. The latter, in shape triangular to polygonal with rounded angles, are however of variable size, the larger ones measuring as much as 7 mm. across. Some of the beams are scarcely Y4 mm. thick at their middle, while others may be 1 mm. and more in width. The majority are more or less flattened in an externo- internal direction, the rest being more or less laterally com- pressed. The nodes are frequently thickened in a knot-like manner or widened into plates of considerable size. As in E.. aspergillam , the entire sieve-plate presents the appearance of being divided by the stronger beams into a number of primary fields and these again subdivided by weaker beams into the in- 66 T. IJIMA : HEXACTINELLIDA. I. dividual meshes. Structurally tlie beams are compact, though the constitueut spicules are never in fusion. On their internal surface are to be seen pores, which are the outlets of excurrent canals ; for, tlie beams possess the chamber-layer, and the inflow and outflow of Avater evidently take place here as well as in the lateral wall. Observations on growing individuals corroborate F. E. Schulze's view ('95, p. 25) with regard to the origin of new beams and meshes in the sieve-plate. These arise by the split- ting and shifting asunder, as it were, of the beams and nodal plates already present. In other words, there arise in the tissues gaps which gradually enlarge into new meshes. I have also seen evidences of new beams, and therewith of new meshes, forming themselves along the inner border of the cufl: Compared with the sieve-plate of U. renalis as figured by F. E. Schulze (19'), that of E. 'wilder ialis presents on the whole much larger meshes. Whereas the greater upper portion of the sponge-wall possesses a certain degree of flexibility and elasticity, the lower portion is firm owing to the fusion of the principal skeletal elements in this region. As already mentioned, the lower end of the body is well-nigh or quite destitute of the loose tissues ; in fact it may be considered as dead. Some distance before the extreme lower end of the skeletal tube is reached, the exposed longitudinal bands of the rather coarse looking, anchoring fibers begin to become frayed out inferiorly, soon to interlock among themselves and to penetrate into the bulbous mud-ball that always makes up the lower termination of the specimens. Not only the basal tuft but also the lowest end of the internal skeletal tube E. IMPEKIALIS. — GEN. CHARACTEKS. 67 itself penetrates the mud-ball. The sponge thus stands tolerably firmly implanted in the substratum, unlike certain other species (e. g., E. marshalli) in which the body, being rooted by the basal tuft only, apparently admits of being subjected to a free sway- ing motion as it stands on the sea-bottom. The buried extremity of the skeletal tube, which is narrowed to about ^/j the diameter of the sieve-plate at the superior end and which is quite dead, is found to be open when cleansed of the mud ; a perforated bottom-plate does not exist. However, in quite young specimens under 75 mm. body-length I have found the inferior end, which probably stands yet unburied in the mud, blindly closed by the living tissues {vide anon). Turning our attention to the features of the parietes on the gastral side (PL II, fig. 5), this surface is as usual checkered with tolerable regularity by the transverse and longitudinal ridges. Much less conspicuous than these are the two systems of the right-handed and the left-handed oblique ridges. All the ridges bear numerous small excurrent apertures, generally not more than Y4 mm. in diameter. Many of the meshes too contain each one large or 2-4 smaller pits, which, by holding the wall against light, can at once be recognized as the apertures of large excurrent canals arising in the external parietal ledges. The rest of the meshes are each occupied by a cup-like or pit- like de- pression, the bottom of which is perforated by a parietal osculum. Not uncommonly two or more of these perforated meshes are found in direct succession either transversely or longitudinally. However, their distribution in relation with that of the other kind of the meshes — the so-called interstitial meshes — must be said on the whole to be iriegular. 68 I. IJIMA : HEXACTINELLIDA. I. The meshes, whether perforate or interstitial, are nearly quadratic but often somewhat elongated in the longitudinal direction. They are usually largest in the middle region of the body, where they may measure as much as 4 mm. by 6 mm. (spec. J of the list given on p. 62). Towards either end of the body they diminish in size down to say 2-3 mm. in length of the sides. The numbers of the transverse and the longitudinal ridges (which correspond to the main skeletal beams pursuing the same direction) as counted on two specimens {D & J oï the list), both of which had fully acquired mature form, were as follows : Spec. D. Spec. .7. Number of transverse beams 112 139 Number of longitudinal beams at the upper end 73(?) 107(?), Ditto, at the middle 47 49 Ditto, at the lower end 27 25 In four more specimens — all macerated skeletons consisting of fused spicules — I have found the number of longitudinal beams at the lower extremity to be 22, 23, 23, and 28 respectively. It then seems that we shall not be wide of the mark in stating generally that the longitudinal beams in old individuals begin at the lower end with a number somewhere between 22 and 28, and that this number nearly doubles at the middle and triples or quadruples at the upper end of the body. It goes without saying that this multiplication is due to the splitting and divergence of the beams in their course. It often happens, especially close to the upper end, that the longitudinal beams are incompletely or but slightly shifted asunder, making it impossible to count their E. IMPERIALIS. — SPICULATION. 69 number with exactness. Lower down on the body, the counting is however comparatively easy, as is also always the case with the transverse beams throughout the entire length of the body, these being everywhere relatively well separated. Spiculation. The parenchymalia prÎ7icipalia are large slender-rayed oxystauractins, in which the longitudinally disposed axis is usually much longer than the transverse. The former is straight and in large specimens may attain a length of nearly 100 mm. and a breadth of 180 /^ or over near the spicular center. The shorter transverse axis may be 30 mm. long. The two lateral rays of this axis are somewhat inwardly directed as they arise from the center, so that they may enter into the composition of the transverse skeletal beams, which are more innerly situated than the longitudinal. Each point of intersection of the two main systems of the skeletal beams is usually, though not always, occupied by a single oxystauractin center ; but not infre- quently it shows none of this. Since therefore the rays are very much longer than the sides of the skeletal meshes, each beam of the skeleton is supported by several oxystauractin rays. Along the inner border of the cuff, i.e., in the uppermost transverse beam of the skeleton, the principalia take a different form in that they are here usually provided with a short distal ray which extends radially into the cuff!', while the superiorly directed, longitudinal ray becomes abortive. The comitalia accompany the rays of the principalia in profusion. They are nearly exclusively elongate thetactins of quite a fine caliber. They may be 30 mm. long with a breadth 70 I. IJIMA : HEXACTINELLIDA. I. of 10-30/^ and more near the spicular center. The rays are for the greater part of nearly uniform thinness ; the ends are slightly swollen and rough, the extreme tip being either rounded or bluntly conical. The unilateral ray may be of considerable length (up to 15 mm.) but more often it is relatively very short, being sometimes only 1 mm. or even less in length. It sticks out of the skeletal beams nearly vertically at indefinite positions and in all directions. Among the comitalia I have on rare occasions met with fine diactins in which the suppressed rays were represented by mere knobs at the center. The longitudinal skeletal beams may attain a thickness of over 1 mm. The transverse beams are on the whole somewhat thinner. The oblique beams of the skeletal framework consist, unlike those of the two other systems, almost entirely of thetactins which are however quite similar to those just described. Some of these thetactins may here be of moderate strength (up to 100 a in thickness) and may be regarded as representing the principalia of the beams. Sometimes such stronger elements were found to be oxydiactins. The spicules of the skeletal beams above referred to begin to undergo synapticular fusion in the well-known manner at the lower end of the body and that at a time when the sponge has acquired a height of about 200 mm. With further growth, the soldering gradually extends upward to about the middle of the body but probably never further than that ; for, even in the largest specimen (X) before me, I find all the spicules in the upper, half in loose association with one another. Here seems to exist another point of difference from the closely allied E. regalis F. E. Sen., in which the soldering j^rocess appears to E. IMPEKIALTS. — SPICULATION. 71 extend right to the upper end oT the bod}^ though not into the sieve-phite. By macerating large specimens and washing away all the free spicules, the wall of the lower half yields a skeletal tube consisting of a continuous, filigree-like latticework of the fused beams (PL II, fig. 9). The tube is narrowest and firmest at the lower end. At its upper end, the beams are frayed out into their separate fibers. The bundles of anchoring spicules, running along the longitudinal beams in the lowest third of the body, do not participate in the fusion except to a very incon- siderable extent in the deepest parts in direct contact with the beams proper. The anchoring spicule or the hasalia (fig. 16) may reach a length of 200 mm. or more and a breadth of 75 [>- at the middle. The axial-cro«s lies at some distance (250-320 A«) from the extreme distal end, which is swollen into the usual miter-shaped knob (75-95 It long, and nearly as broad) furnished with a whorl of 5-9 anchor- teeth. The latter are much smaller and shorter than in E. marshalli or oiveni. Soon after its origin from the knob, the shaft is only 19-23 /j. thick. The first barb-like spine, on following the spicule from the distal end, occurs shortly in front of or behind the position of the axial-cross. I observe no definite rule as to the arrangement of the spines on the shaft. They may extend proximally for nearly half the length or more of the entire spicule, imparting to that portion a peculiarly glistening appearance when seen with the naked eye. Proximally they become gradually smaller and wider apart until they altoge- ther cease to exist, leaving the rest of the spicule perfectly smooth up to the finely attenuated upper end. Intermixed among the bundles of basalia-shafts on the sponge-wall, there are always 72 T. IJIMA : HEXACTINELLIDA. I. found, but more frequently in small than in old specimens, short and young anchoring needles in various stages of development. The anchor-heads attain full size in this position and are sub- sequently, along with continued elongation of the shaft, pushed on downwards, finally to penetrate into the substratum. The parenchymalia which support the loose tissue covering up the skeletal latticework, are again chiefly thetactins with short unilateral rays ; but not uncommonly they are also hexactins, usually with one axis which to a greater or less degree is more elongated than the others ; and occasionally they are diactins, generally with knob-like indications of suppressed rays at the spicular center. Besides these, there may occur, though ex- ceptionally, any other form of spicules. All are small to medium- sized spicules, ^'ji-l mm. in length and with rays 15-30/^ thick near the center. The ends of the rays are sparingly beset with spinules and usually terminate in a conical point. These puren- chymalia frequently combine into loose, ill-defined strands, which mostly extend peripherad ; otherwise, they stand isolated either without any order of arrangement or with one axis pointing towards the surface. Among the diactin-parenchynialia, those that are exlernally protruded as jyrostalia at the edge of the cuff and of certain parietal lappets, require special mention. AVe have here to do with slender oxydiactins of very variable size — up to 5 mm. in length and 20/^ in breadth — in which the center is indicated by a slight annular swelling. In the positions indicated, such oxy- diactins are usually numerously pre-ent and disposed in radial arrangement. Some lie still completely imbedded in the body- wall ; others are partially or completely projected out of the E. IMPERIALIS. — SPICULATION. 73 surface in coherent bundles, wliicli stand each in association with the distal ray of certain especially strongly developed dermal hexactins presently to be noticed. The bundles constitute the bristle-like prostalia already described. Just the same kind of prostalia is known to occur also in some other species ; e.g., in E. marshal! i Ij., JE. nodosa F. E. Sch. The osoularia (fig. 17) are mostly thick-rayed, plump- looking hexactins and pentactins. Both these forms occur in about the same numerical proportion. In addition to them, there also occur not uncommonly forms which nearly approach or virtually are stauractins or compass- needle-like diactins. Their size is exceedingly variable, those near the edge of the oscular membrane being much smaller than others situated in more peripheral positions in the zone. The former may measure only 120/^ in axial length and lb !> in breadth of the rays, while some of the latter may attain 4öO /^ and 50/^ in the corresponding dimensions. The rays are either rounded at the end or pointed and conical in shape. The spicules lie thickly crowded for the most part in several layers and apparently without regularity as to the disposition of the axes in relation to the surface of the zone. In the case of oscula of comparatively recent formation, the oscularia are always much less numerous than in those of long s'tanding. Noteworthy is the fact that there exist in the immediate neighborhood of the zone certain hexactins and pen- tactins, which, considered in respect of position, size and general appearance, might well be considered as standing intermediately between the oscularia on the one hand and the hexactin-paren- chymalia and the pentactin-gastralia on the other. 74 I. IJIMA : HEXACTINELLIDA. I. The sword-shaped oxyhexactin dermalia of the general surface are comparative!}^ small and slender, the ray usually measuring less than 1.5 mm. in length and 7-12,« in breadth near the spicular center. The distal hilt- ray is only 130-180/'- long. Exactly as in E. regain, it gradually tapers distally to a fine or conical point and shows obsolete prickles which stand closely together near the outer end but are isolated and sparingly present on the rest of the ray. The five remaining rays are rough only at the ends. The paratangentials (220-350 /i long) of different dermalia tend to form a rectangular mesh work (meshes about 3 mm. wide) in the depressed areas of the external surface ; towards the summits or edges of the parietal prominences the arrangement becomes irregular. The blade-ray, which pierces the choanosome like a nail, is usually several times longer than the hilt-ray, — occasionally only twice but more often it is nearly ten times as long. Unusually large and strong hexactin-dermalia occur, together with others of the more ordinary dimensions, along the cuff-edge as well as on the highest parts of the parietal ledges, especially in conjunction with the bristle-like prostalia already mentioned. In these positions they may attain a size more than thrice as large as the ordinary dermalia. In one specimen measured the greatest axial length was 4.7 mm., of wdiich 1.2 mm. belonged to the hilt- ray, the breadth of the rays near the center being 40^. 'While on the one hand there exist intermediate transitional forms between the large and the small dermalia, some of the former are, on the other hand, more or less deeply situated below the others, so that they appear sometimes not unlike hypodermalia or otherwise assume such positions as seem to justify their being taken for parenchymal oxyhexactins. E. IMPERIALIS. — SPICULATION. 75 The gastralia are oxypeutactins sparingly supplied with minute prickles at tlie conically pointed ends of rays. The paratangentials (200-450/^ long, 13-15/^ thick) are frequently of unequal length in the same spicule and more or less bent so as to form an irregular cross. The unpaired, distally directed ray is straight and somewhat longer. As in all other species of the genus, the gastralia are nowhere so regularly arranged as to form a quadrate meshed latticework. Quite similar pentactins extend into the excurrent canals as canalaria, which become more and more sparse toward the distal end of the canals. The floricomes look exactly like those of other Euplectella. In diameter they measure 90-105 p., say about 97 /'- on the average. Only in very young specimens have I found them perceptibly smaller (84-91 /-« dia.), but never so small as in E. marshaUL The number of terminals to each principal ray varies from 7 to 12. The terminal plate shows 5-9 sharp teeth on the external edge, while its internal edge is represented by a simple obtuse rounding of the surface, as is usual with all floricomes (PI. II, fig. 14d). In sections of a specimen 210 mm. long, I have found in abundance cases of the floricome in various stages of developing its terminals (figs. 10-12, 14). They were all situated in the subdermal trabecular space, which undoubtedly is the place where the rosette in question arises and reaches full development, eventually to be moved off to the apex of the hilt-ray of the dermalia. In the earliest stage observed, the six princi^^als were already fully developed though still somewhat thinner than in the mature state. Each principal, traversed throughout by the axial canal, terminated externally in a lenticular disc, from the 76 I. IJDIA : HEXACÏINELLIDA. I. margin of which arose a numher of short, and uniformly exceedingly fine terminals in a single whorl. The terminals measured scarcely over 10/^ in length and together formed a wine-glass-like perianth (fig. 10). I have failed to discover a still younger stage, much as I have wished to do so ; so tliat, the mode of development of the principals must ever remain entirely in the dark. On the other hand, I have succeeded in ohserving a continuous gradation- al series of forms leading from the above-described stage up to the completely developed floricome. The fine terminals elongate and by flaring out at the outer end convert the perianth into a deep bell-like shape (sigmatocome, fig. 11). (See p. 52). The outer portion of the terminals has somewhat thickened, but the extreme tip appears still pointed (fig. 14a). It continues to thicken especially at the tip ; meanwhile, the latter passes into a state which, when observed under a high power, appears as obliquely and somewhat roundly truncated (fig. 12; fig. 14*^). It may now show small rudiments of the marginal teeth (fig. 14c) and indis- putably presents itself as an inceptional terminal plate of the floricome. Fig. 14 d represents a fully developed terminal in the same scale of magnification as the developmental stages a, b, and c of the same figure. In E. aspera, F. E. Schulze ('95, p. 29) found the terminals of young floricomes with a knob-like swelling at the free end. In E. wiperialis and E. marshalli, this is never exactly knob- like but rather obliquely truncated as already mentioned. The oxj/hexasters (fig. 15) measure 83-92 /i, mostly about 86/^, in diameter. They are by far less numerous than the floricomes. While in some places in the deeper parts several oxyhexasters were found together at no great distances from one another, they E. IMPERIALIS. — SPICULATION. 77 were decidedly rare in the parietal ledges. The principals are exce.-sively short, being represented by a knob-like swelling separated from the central node by a narrow constriction. They are quite unlike those in E. regcdis, in which, according to F. E. Schulze's representation, they should be short but narrow. The strongly divergent terminals, 3-5 to each principal, are smooth, nearly straight and of moderate length. The graphiocomes measure up to 330/.« in diameter. They are tolerably common everywhere in the periphery of the wall, though it may be comparatively rare to meet with one in a perfectly intact state. The sheaves of terminals may be 154/^ long and 20/'- broad. Detached and isolated terminals — i. e., the rhaphides — are scattered here and there in the superficial region, lying irregularly but mostly more or less vertically to the surface along with the hilt- rays of the dermal ia. In comparison wdth E. marshalli, the rhaphides so situated are not so numerous. The central remnant of the rosette after completely shedding off the rhaphidial terminals, has been very frequently met with. The discs at the end of principals are then seen studded all over their external surface with small prickles. Different stages of the srowth of the terminal sheaf have also been found. The rhaphides composing each sheaf are at first very short and ex- ceedingly fine (fig. 13). For further account of their develop- ment, see under E. marshalli. In E. regalis, which clearly is very nearly related to the present species or to E. aspergillum, F. E. Schulze has entirely missed the graphiocome; nor does he appear to have seen any free rhaphides. Nevertheless, I consider it not altogether im- 78 I. IJIMA : HEXACTINELLIDA. I. possible that further research with more materials may reveal the presence of the rosette in that species also. The sieve-plate presents a spiculation somewhat differentiated from that of the lateral wall. The principal parenchymalia are here oxydiactins, smooth at the center and gradually attenuating toward either end. They may attain 30 mm. or over in length and 190/^ in breadth at the center. They are usually more or less curved and often rather abruptly bent in accommodation to the corners of the sieve-plate meshes. The accessoria, copiously present and forming close bundles with the principalia, are likewise predominantly diactins and occasionally thetactins. The diactin forms show either annulated or cruciately tubercled centers. Certain diactins are remarkably short in relation to their thick- ness and may be called compass-needle-like. The external surface of the sieve-plate exhibits nearly regular hexactin-dermalia of 150-400/^ axial length, sparingly present on the thinner beams but densely crowded on the nodal plates. The rays in these spicules are relatively strong and acutely or bluntly pointed at the roughened ends ; the proximal ray i^ierces right into the sub- jacent parenchymal bundle. The gastralia on the inner surface are less abundant. They are pentactins of an appearance quite similar to the dermalia save the absence of the free ray. All the three forms of rosettes found in the lateral wall occur in the sieve-plate also, — the floricome and the oxyhexaster very sparingly but the graphiocome in abundance. Hence, free rhaphides are of quite common occurrence in the dermal layer. Finally, it may here be added that I have ascertained by sections the presence of the chamber-layer in the sieve-plate. E. IMPERIALTS. — YOUNG SPECIMENS. 79 Young Specbiens. Of quite young E. imperialis, under 75 mm. length (excl. of the basal tuft), I have been able to bring together no less than a dozen specimens, of wliicli the smallest (^Pl. II, fig. 6) measures only 30 mm. in body-length and 10 mm. in greatest breadth. They may be described in general as follows : The body is bellied in a spindle-like manner, straight or slightly bent and circular in cross-section. Inferiorly it narrows to a conically closed end, whence arise the basal spicules in a small, almost solid tuft. Above, the body contracts in a much less degree and ends almost truncated, the sieve-plate being only slightly convex. The latter structure is very frail, the beams being quite thin ; the angular meshes number from about half a dozen to a score according to the size of the specimens. There is yet scarcely a trace of the cuff. Numerous small parietal oscula occur already in the smallest specimen above referred to. The general form thus closely resembles that of young E. mar- shalU, in fact I think of all ^w^/ec/e//a species in the correspond- ing stage of growth. However, there are certain points by which the similarly sized young of E. imperialis and E. marshalli may be distinguished. Firstly, in the former the external surface presents a more jagged appearance and shows a larger number of small, bristle-like, prostal spicules, while in the latter the broader and more con- tinuous parietal ledges present on the whole a nearly even surface (c/r. PI. II, figs. 6 & 7 with PI. IV, figs. 8 & 9). After attaining a body-length of about 70 mm., the parietal ledges in E. imperialis are already developed into the characteristic lappets or interrupted 80 I. IJIMA : HEX ACTIN ELLIDA. I. ridges, which, in such young specimens much more frequently than in the older ones, are fringed with an irregular row of bristle-like prostalia. Fig. 3, PL I, may be taken as an illustra- tion of the general appearance of the sponge at the stage in question. The parietal prominences are best developed on the middle of the body but are yet quite undeveloped close to the sieve-plate end. They are nowhere so conspicuous as in old specimens, but when compared with E. marshalli of about the same size, they give a jagged appearance to the sponge surface sufficient to serve as a distinguishing feature. Secondly, the size of the floricome presents a certain con- stant difference in the two species. It seems that in all Euplectella species the said rosette is on an average somewhat smaller in young than in old specimens. Now, whereas in old E. marshalli it never exceeds 80 /-« in diameter, all the small E. imjyerialis under consideration have it appreciably larger ; here the diameter may reach 91 /^ or over, although some other floricomes in the same individual may run down to 84 // in diameter. Further there are some other points w^iicli at times may serve as an aid in referring young specimens to one or the other of the species in question. 1) The locality and depth from which the specimens were obtained ; for, so far as my knowledge goes, tlie two species seem to have each its own sphere of distribution both horizontally and vertically (see p. 60). 2) The character of the substratum as exemplified by the matter contained in the root-tuft; for, while E. imperialis exclusively inhabits sandy or muddy bottoms, the other species is apparently confined to the coarser shelly grounds. 3) Tlie species of the Crustacean immate (see anon, under Miscellaneous Notes). Constant as the difference in this regard seems to be, the drawback is that in E. IMPERIALIS. — YOUNG SPECIMENS. 81 very young EupIecteUa the inmate is more often absent than present. I may here add that the sieve-plate of young E. imperialis is perhaps less frail and less liable to be lost than in E. marshalli of a similarly small size ; for, I have found it preserved intact in most cases of the former, while it was broken and lost in the majority of the latter. Further I may record that whereas in young E. imperialis the sieve-plate was always only flatly convex, it was often, though not always, much more prominently so in individuals of the other species in nearly the same stage of growth. The approximate numbers of transverse and longitudinal beams of the skeleton, as counted on four small specimens of E. imperialis, were as follows : Spec. Size of body. Number of transverse beams. Number of longitudinal beams at middle of body. Length (excl. of basal tuftj. Greatest breadth. a. h. c. d. mm. 32 34 48 60 mm. 8 8 12 11 30 2.5 26 + 40 26 20(?) 23 22 A comparison with the numbers of corresponding beams in fully adult specimens as given on p. 68, will at once show that those of the longitudinal beams in the young (last row of the above table) are, generally speaking, nearly equal to, or at any rate not widely at variance with, the same as counted at the lower end of mature specimens. This should mean that in the basal region the longitudinal beams develop to their full or nearly 82 I. IJIMA : HEXACTINELLIDA. I. full number at a very early period of life. With the growth in girth of the sponge, the beams in question become wider and wider set apart from one another, and if increase in their number takes place at all in the said region, it can not be to any con- siderable extent. An end is definitely put to their shifting asunder or their new formation by the soldering together of spicules, which process, as already mentioned, sets in at the sponge-base when the entire body has attained a length of about 200 mm. Quite a dififerent circumstance should obtain in the upper region. With the growth of the body, many of the ever elongating longitudinal beams undergo splitting at indefinite positions in the circumference but especially frequently near the upper end to- wards the close of the growth, when that end becomes the broadest part of the entire body. With respect to the increase in number of the transverse beams, the facts ascertained by F. E. Schulze ('95, p. 2")) from E. simplex find corroboration in the present species. The most active seat of their multiplication is the upper end of the sponge close to the cuff, where they lie most closely together and are thinnest, being composed of slender and evidently young parenchymalia. Here the sponge-wall is youngest at all stages of its growth and the formation of new transverse beams may be said to be constantly taking place until the sponge has nearly reached its full length. After the expansion of the upper end into the broadest part of the wall, there no longer exist signs of their new formation. The lower extremity of the body proper is, as already re- marked, blindly closed at first and apparently lies free above the surface of the substratum, the sponge being rooted by the basal spicules only. Soon the condition changes. By the time E. IMPERIALIS. — MISCELLANEOUS NOTES. 83 the sponge measures 75-100 mm. in length, the lowest end of the body encroaches upon and begins to bury itself in the substratum. This seems to cause degeneration of the soft tissues at the extreme inferior end, which henceforth remains open as before described. A perforated bottom-plate, such as occurs in E. marfihaUi, JE. oweni, E. jovis, &c., does not here come into formation ; it develops, in my opinion, only in those species in which the lower end of the body remains life-long apart from the substratum. Miscellaneous Notes. E. iinperialishas recently been pointed out by F. E. Schulze (19', p. 29) as having its nearest ally in E. regalis of the Indian Ocean. I should think the latter species is about as nearly, if not more nearly, related to E. ai^pergUlum. Indeed, so close seems the agreement between the single specimen on which E. regalis is based and the not fully mature specimens of the Philippine species, that, should the graphiocome come to be dis- covered in the Indian form — which contingency I presume to be not altogether unlikely with examination of more specimens — there w^ould remain probably only some slight differences in ex- ternal form to fall back upon as differential characters between the two species. On several s^^ecimens of E. imperialis were observed un- mistakeable signs of the regeneration of loose tissues at such parts of the external surface as had been stripped of the ledges and flake-tissues by some mechanical cause. Far more substan- tial deformities arise in connection with the repairing of such injuries as the breaking ofi', rending, twisting or bending of the 84 I. IJIMA : HEXACTINELLIDA. I. body- wall. Remarkable was tbe case of a specimen of medium size in whicb the body, instead of being tubular, was more like a compressed sac, — probably brought about in consequence of tearing and other injuries sustained by the wall at several points. The upper end of this specimen was closed by a sieve-plate wliich bore the appearance of having been secondarily formed after the loss of the original one. It would be superfluous to enumerate all the other cases of malformations due to similar causes. Once a specimen was obtained which was normal in all respects except in bearing on one side near the upper end a second shorter tube, connected with the first by means of a short, solid, lateral bridge, through which the parenchymalia of the one tube passed into the other. The smaller tube thus appended looked very much like the upper torn end of a distinct indivi- dual, having a regular sieve- plate above but being closed by regenerated tissues at the opposite end. In fact I do not know how to explain this abnormality unless it be that we really have to do with such a portion of a separate individual which came into fusion with its neighbor previous to its becoming cut off from its basal region. Finally a few words about the commensal inmates. It is interesting to note that the two species of Euplectella occurring in the Sagami Sea, viz., E. imjwrlalis and E. 7/iarshaIli, have each a special macrurous Crustacea as inmate in the gastral cavity. For the first named species, this was a species which unfortunately could not be satisfactorily identified, although Dr. KisHiNOUYE for my sake kindly made efforts at its deter- mination. It conies nearest to both Alpheidte and Ilyppolyt- E. IMPERIALIS. — MISCELLANEOUS NOTES. 85 idse and yet is not exactly admissible into either family. The body of the Crustacea, which may be about 40 ram. long, is laterally compressed ; the keeled cephalou is bent in hump- like manner and suj^plied with a small rostrum ; the eyes are exposed, not covered by the prolongation of the cephalic cara- pace ; the mandibles with palpi ; the third maxilliped robustly developed ; the first leg chelate and symmetrically paired ; etc. A noteworthy circumstance is the fact that this Crustacea occurs each time singly in the gastral cavity — not in a pair as is usually the case with Spo7igicola venusta of E. marshalli and E. oioenL The Crustacea was only exceptionally absent in the larger specimens of the sponge but quite frequently in the smaller under 75 mm. body-length. A certain relation seems to exist between the size of the inmate and that of the host, probably as the result of the former entering the latter when both are yet small and their continuing to grow together. Once an oxyrhynchous crab, Chor ilia (closely resembling C, longipes Dana), was found instead of the Macroura. Among the other inmates Ophiurons were not uncommonly represented by two or three different species at a time. 86 I. IJIMA : HEXACTINEELIDA. I. EUPLECTELLA MARSHALLI Ij. Pis. iir, IV & V. EvplecteUa JIarsIialli, Ijima, '95, p. 93. Under the above designation in 1895 I briefly described a second species of Euplectella inhabiting the Sagami Sea. Though evidently a very close relative of E. owenl Herkl. & Marsh,, long known as from Japan, it clearly deserves to be considered as a distinct species. E. marshaUi became originally known to me from Döketsba (see p. 13). Depth 7-5-160 fms. (137-292 m.) ; bottom shelly. As spots where one may be almost certain of securing specimens of the species, I may mention : Matswa Lighthouse line by 3Iera 2-3 ; Ena line by Mera 2-3 ; cfec. Since the first specimen was brought to me by Kuma in 1894, scores must have been taken on the same ground by different collectors, I myself having to account for a goodly number. Apparently the species flourishes there in luxuriance, alone by itself and not in association with E. imperialis, which belongs to a deeper bottom of a quite different nature. It has thus afforded me a rich series of materials which has enabled me to institute a much closer study of it than of any other species of the Hexactinellida. However, Doketsba does not seem to be the only locality where E. marshaUi is found. During the cruise of S.S. * Onoura Maru ' on behalf of the Fishery Bureau of the Japanese govern- ment in 1893, a young specimen, only 44 mm. long, was obtained off Okada, a small village on the northern coast of EUPLECTELLA MARSHALLI. 87 Vries Island. Nothing further is known about the circumstances of its capture. The specimen was kindly given me by Dr. KiSHixouYE for investigation and I hold it as belonging to the species now under consideration. It contained in the gastral cavity a single Si:)ong\cola venusta. Further on May 7, 1900, the U. S. Fish Commission Steamer ' Albatross ' obtained by means of tangles a small, much macerated Euplectella at her Station 3700 (Senoumi in Suruga Gulf; depth only 73 fms. [133 m.]; bottom volcanic mud and sand). I was on board the ship at the time and although I could not undertake a microscopical examination of the specimen, I have judged it to belong to E. marshalli from its general external appearance. The Crustacean inmate was Spongicola venusia. The other catches at the same station were quite similar to those usually obtained at Döketsba. The general color of the sponge in the fresh state may be called a light salmon or a pinkish buff. It is deeper and brighter than in E. imperialis. When dried or preserved in alcohol, specimens become entirely colorless. In fresh specimens obtained by me July 17, 1895 and April 1, 1900, I have noticed with the naked eye numerous small spots of a deep orange-yellow in the substance of the sponge-wall. The larger spots, irregular in outline, measured nearly half a millimeter across. They were most distinctly visible on the gastral side but were also present imbedded in the deeper parts of the parietal ledges. I believe the same spots are to be seen in a greater or less quantity in all individuals and at all seasons of the year. When put into alcohol, the orange-yellow color is dissolved away and the spots become lost in the general white- 88 I. IJIMA : HEXACTINELLIDA. I. ness of the tissues. In specimens preserved in weak formalin, the color has likewise disappeared but the spots remain distin- guishable, being more opaque than the surrounding tissues. Osmic acid has blackened the spots. I think the coloring matter referred to, which is combined with oil-like spherules contained in the cells composing the spots, is the same as that which is diffused in the soft tissues and gives to these the before men- tioned pinkish-buff color. The cells of the spots will find further treatment under the histology of the soft parts. (See anon under Thesocytes). General Characters of Nearly or Quite Full-Grown Specimens. The body may be described as a gently bellied tube with an irregularly corrugated external surface (PI. III). The sieve- plate is strongly arched ; the basal tuft, large and elongate. The broadest part of the body is usually, but not always, situated some- what below the middle; so that, the general shape is frequently not unlike that of a lamp-chimney, while in others it more resembles a barrel or a cucumber. Young specimens show a shape approaching that of a spindle. The cross-section is on the whole circular, except at the extreme lower end which is as a rule more or less distinctly compressed. The following are measurements of some of the larger specimens taken up at random : E. MARSHALLI. — GEN. CHARACTERS. 80 Spec. Length of body (excl. of basal tuft). Diameter of sieve-plate. Diameter of body at tlie broadest part (external ledges included)*. Diametei-s at the compressed lower end of body. A mm. 122 mm. 25 mm. 44-.54 mm: 23-27 B 122 30-33 45-53 14-25 0 143 32 5.5-59 17-31 D 144 21-29 44-52 12-14 E 155 28 46-48 12-16 F 182 41 63-65 10-34 G 193 35 60 1.5-37 Specimen G of the above list is about the largest specimen of the species I have as yet met with. The body-wall measures not more than 3 mm, in thickness, leaving the height of the parietal ledges out of consideration. The ratio of the body-length to the greatest breadth may be put down as 1 : 0.3-0.44. In comparison with either E. imperialis or E. oweni, the body is considerably shorter in relation to its breadth. The parietal oscula, not over 2 mm. in diameter, are rather irregularly scattered. They lie in broadly pit-like or elongate valley-like depressions of the external surface, sometimes singly and at other times in groups, but without a definite rule as to their relative position. Those opening on the same depressed area are separated from one another by an interspace which is either gently convex or nearly flat and varies from 1 mm. to *The fluctuation of diameter in the same individual, to be noticed in this column, is mainly due to the various degrees of tlie development of the parietal ledges at different points of the circumference. Leaving aside these ledges, the cross-section of the body may be said in general to be approximately circular. 90 I. IJIMA : HEXACTINELLIDA. I. 5 mm. in width. Others separated by a prominent parietal ledge are frequently 8 mm. or more apart. Th.Q ixirietal ledge is very well developed and gives a strikingly characteristic appearance to the sponge. It is broad and generally round-edged but quite irregular as to tlie course and configura- tion it takes. Sometimes it remains rather low and exhibits an approach to a checkered arrangement in that it runs in inter- secting transverse and longitudinal systems (see fig. 2, PL III). Much more usually the ledges rise in irregular crests, lappets or tubercles of variable height and extent. These present an appearance on the whole quite different from those of E. imjjerialls in being broader and more bold in their outlines. The crests may pursue a plainly oblique course after the manner of those in E. asiiergillwn (fig. 1, PL III), but this is by no means general. In fact, they may run in any direction, often tortuously and as often branching and anastomosing in an altogether in- definite manner. They are generally in greatest development where the sponge-body is broadest. Here they may be 10 mm., sometimes even 15 mm., high, as measured above the level of the parietal oscula. Close to either end of the body, the ledges are on the whole low, though by no means uniformly so at different points or in different individuals. Originally separate parts of adjacent ledges may, during growth, come into contact and fuse together. In this way is to be explained the origin of the bridge-like connections which have been now and^then observed. The growing ledge may occasion- ally so encroach from all sides upon the position of a parietal oscula, that there finally arises a narrow canal opening externally E. MAESHALLT. — GEN. CHARACTERS. 91 at some point on tlie ledge itself and leading internally into the gastral cavity. The fine dermal laiticeivork is scarcely discernible by the naked eye but under a hand-lens may be seen to extend all over the external surface, except on the thin iris-like oscular membrane. The nodes of the delicate lattice generally appear as minute whitish spots. Through the layer are everywhere to be seen, in varying degrees of distinctness, the entrances into the incurrent canals. These are of all sizes under 1 mm. diameter. A cuff is usually more or less distinctly present. In several specimens, it was plainly developed for only a portion of the sieve-plate circumference and merely suggested for the rest. It is as a rule directed obliquely upwards ; exceptionally, however, it becomes nearly horizontal. Its free edge is always uneven and and occasionally even deeply indented, thus giving unequal breadth to the cuff in different parts. Measured on the upper surface, the breadth may be as much as 10 mm., but generally is much less. On the lower side, the cuff is joined by the parietal ledges which run between the last oscula at this end and are here comparatively low. Besides being somewhat thinner, the cuff differs from the average crests of the parietal ledge in being rather sharp-edged instead of being rounded, and in showing a fringe of fine, inconspicuous marginalia, projecting to a length of less than 1 mm. But this difference is of no essential im- portance, since certain parietal crests have been occasionally found having characteristics exactly like those of the cuff. The sieve-plate may show in some cases a convexity nearly 92 I. IJIMA : HEXACTINELLIDA. I. equal to that of an ordinary watch-glass. More frequently it presents a stronger arching, being often of the shape of a hemi- spherical vault. The meshes are smaller than in E. imperialis and never exceed 4 mm. in their greatest width ; they have mostly a shape varying from oval to triangular or polygonal, the corners in the latter cases being invariably rounded. The beams are comparatively strong-looking and are distinctly laterally com- pressed, like those of E. aspergillwn or E. oiveni. Seen from above, the majority are less than 1 mm. or even half a millimeter in width ; but the same beams, when seen from the sides, may be considerably wider, up to nearly 2 mm. in the case of a strong beam. Towards the nodes, as seen from above, the beams either maintain their width uniformly or broaden so as to form a more or less distinct nodal thickening. Occasionally a beam or a node is so broad (up to 3 mm.) as to deserve to be called plate-like. A distinction of tlie meshes into the greater and the lesser — the former bounded by wider beams and containing a number of the latter — can not be made in the present species. The surface of the sieve- plate beams is on the external side close-grained and compact-looking, while on the inner side are seen small excurrent openings scattered between separate strands of fibers. The lower end of the sponge-body (PI. IV, fig. 5) is likewise closed by a perforated plate, the bottom-plaie (see p. 40). Though often found in a damaged condition, the occurrence of this plate seems to be constant in the present species. It is a direct conti- nuation of the lateral parietes ; thin, measuring not more than about 172 niïn. in thickness ; tolerably even on both surfaces ; and nearly flat or outwardly convex. The skeletal framework E. MAKSHALLI. — GEN. CHHRACTEES. Ö3 of the lateral wall ceases to exist just arouud the bottom-plate, so that the texture of the latter is entirely soft, being nearly exactly like that of the parietal ledges. There can be no doubt whatever that the round perforations of the plate are morpholo- gically and functionally the same as the parietal oscula or the sieve-plate meshes. In the main middle portion of the bottom- plate, the oscula are irregularly scattered, while in the periphery directly adjoining the last circular beam of the skeletal frame- work, they are somewhat larger and more closely set, leaving between them narrow beams by which the more central portion of the plate is attached to the lateral wall. The plate therefore most easily breaks off at these weak points in the periphery. The internal surface shows a number of small excurrent aper- tures ; the external surface appears essentially the same as that of the lateral wall. The bundles of basal anchoring spicules emerge from the lateral wall in a circle near the juncture of the latter with the bottom-plate, soon to form a soft and silky basal tuft in the usual manner. Since now, as already noticed, the lower end of the sponge-body is compressed, the inverted hollow cone formed by the basal spicules immediately after their emanation from the wall, is likewise laterally flattened. More properly speaking, this primal portion of the basal tuft is wedge-like in shape. It is perfectly free of foreign objects interlocked among the fibers and I have no doubt that this portion stands above and clear of the surface of the substratum. Now, the compressed state of the tuft just before it strikes root into the substratum, would give to the sponge a greater freedom for swaying in one direction than in any other. I think it quite likely that this circumstance stands in a definite 94 I. IJIMA : HEXACTINELLIDA. I. relation to the prevailing direction of the movement of the sur- rounding medium. The lateral compression of the body, especially at the base, in so many other Lyssacine Hexactinellids may possibly fall under the same category of phenomena. For the rest the basal tuft penetrates into the sea-bottom and represents an irreguhxr, elongate and often bulky mass, in- closing a fair sample of the bottom (fragments of sponges, Bryozoa, Mollusca, &c.; sand and pebbles). The lump is fiequentl}'' much longer than the sponge-body proper. I can not tell whether or not, it is simply due to the loose character of the substratum in this case, that such a large basal mass is taken up with the sponge. The appearance of the wall on the gastral side (PI. IV, fig. 4) is essentially the same as in other Eupledella. The transverse ridges are on tlie whole well set apart from one another, notwithstanding the fi equent occurrence of anastomoses by means of ohliquely running ridges. The longitudinal ridges show less regularity of arrangement in so far as they often run in pairs unusually close together. Such double ridges occur without definiteness as to their position. The two may remain nearly unchanged in relative position throughout their entire length ; at other times they may in their course either fuse into one or diverge into two indubitably distinct ridges. Close to their juncture with the sieve-plate, the transverse ridges are usually scarcely recognizable as such. We rather see here only obliquely and longitudinally running ridges or beams, which above pass directly into the formation of sieve-plate beams (upper part of fig. 4). The beams of the skeletal framework, which form all the E. MAESHALLI. — GEÎT. CHARACTERS. 95 ridges seen on the gastral surface, are more often composed of several strands of fibers in close union, than of a single compact strand. With respect to the relation between the different sys- tems of skeletal beams, it may here be mentioned that the oblique beams very frequently blend into the transverse as well as into the longitudinal beams, besides entering into all sorts of relations at the points where they intersect these. Some oblique beams are plainly seen to pass between these two systems ; some to penetrate through their separate strands or fibers ; and others to pass over even to the inside of the transverse or to the out- side of the longitudinal beams. In short, the oblique system of skeletal beams may be said to permeate the two other systems rather than to occupy an intermediate position. As a separate system, the oblique beams in E. marshaUi are apparently in a state of less differentiation than in many other species, e. g., E. imperialis. I shall have later to return to the adult skeletal beams when I come to treat of quite young specimens. The rectangular meshes, formed by intersecting transverse and longitudinal ridges on tlie gastral surface, are either quad- ratic or somewhat elongated in the longitudinal direction. They are largest at the most out-bulged portion of the sponge, where they may measure 4-6 mm. in length of sides in large specimens. In conformity with the irregular distribution of the parietal oscula as seen from the outside, no definite rule can be laid down with regard to the relative distribution of the so-called interstitial meshes and those perforated by the oscula. Not infrequently, either kind of mesh may be seen several in succession transversely, longitudinally or obliquely in one direction or the other. 96 i. ijima : hexactinellida. t. Spiculation. E. Marshalli is one of those species whose spicules remain all perfectly free, never and nowhere becoming soldered together by synapticula?. I am inclined to bring this in correlation with the fact that no point of the living parts is in direct contact with the substratum (see p. 45). The composition of the ^9a?'e«c7i^?/w?«/i« is essentially the same as that which I have described for E. imperialis. The large or medium-sized stauractin principalla, common to the transverse and the longitudinal beams of the skeletal framework, have slender tapering rays which are subterminally roughened and generally conically pointed at the ends. The rays may be 110/^ thick near the spiciilar center, the longitudinal axis being 4o mm. or more and the transverse 20 mm. in length. The two rays constituting the former axis are usually of unequal length but form a straight line, w^hile those of the latter are symmetrical in length and form with each other an obtuse angle open towards the axis of the body, or are at any rate bent to conform with the curvature of the circumference. Toward either end of the sponge, the stauractins become smaller, the rays at the same time approximating in their relative length. The comitalia to the above principalia are, as usual, mostly thetactins ; occasionally also hexactins, paratetractins or pentactins, and rarely diactins. In all these, the rays in an axis, which runs along with the principnlia, are greatly prolonged (up to 2 mm. or more in axial length) in excess over the remaining ray or rays -which spring out more or less vertically from the beams. The diactin comitalia have the center marked by an E. MARSHALLT. — SPICULATION. 97 annular swelling or by criiciately disposed knob-like rudiments of the suppressed rays. The prolonged comital rays are mostly under 25/^ in breadth near the center; they taper outwards, thence to maintain a nearly uniformly thick, filamentous caliber to the end, which is rounded or conically pointed and subter- minally rough-surfaced. Scattered here and there among the fibers of the beams in question, I have on certain occasions found small hexactins, pentactins, stauractins and such like, whose short rays made them appear to be somewhat distinct from the other much elongated elements but which are probably to be classed together with these simply as cases of arrested development. The oblique beams of the skeleton show a similar composition save the absence of stauractin-principalia. Slender thetactins and paratetractins predominate among their elements ; frequently intermixed with these are pentactins and hexactins of moderate strength. Just the same elements constitute the parenchymalia of the flake-tissue, in which they are arranged either loosely or in small bundles. A number of the latter in radial and rafter- like arrangement serve to support the parietal ledge, similarly as described by F. E. Schulze ('87) for^^. aspergillum. As a category of spicules closely associated with parenchy- malia must be considered the prosfal oxi/diadlns, which, occurring in comitalia-like bundles around the distal rays of certain der- malia, cause inconspicuous briitle-like projections along the edge of the cuff and of certain parietal crests. The same sort of spicules is also known in JiJ. imperlalis, E. nodosa, &c. (p. 72). In the present species, the oxydiactins in^ question are small, being at most about 1 mm. long and not exceeding 8/^ in thick- 98 I. IJIMA : HEXACTINELLIDA. T. ness near the center. They are smooth all over but usually show a gentle annular swelling at the center. They are not to to be confounded with the rhaphides (PI. V, fig. 36) which are similarly grouped around the distal rays of dermalia but are of a quite different origin and character. The oscularia (PL IV, fig. 27) are of quite varied shapes and sizes. The more common forms are diactins (compass-needle- like, with either two oppositely or four cruciately disposed central knobs), thetactins, stauractins, paratetractins and pentactins. Rays smooth, moderately thick, reaching 330 ;'- or more in lenoth and 35/'- in thickness near the center. It is difiicult to jwint out which of the above mentioned forms is the predominant. The diactins are more commonly located near tlie edge of the oscular membrane ; some of the thetactin and pentactin forms stand intermediate respectively to the parenchymalia and the gastralia in points of general appearance and mode of occurrence. For differences from the oscularia of E. impevialis and E. oweni, compare fig. 17, PI. IF, with fig. 10, PI. VI. The basalia (PL IV, fig. 26) have a broadly miter-shaped anchor-head, measuring 90-110// across from tip to tip of the oppositely standing teeth. The latter are strongly developed and are 5 or 6, sometimes 7, in number. The apex of the head is either rounded or pointed as in a Gothic arch. The shaft is about 26/^ thick at the point of its origin from the head, whence it gradually narrows above until at a short distance above the position of the axial cross (which lies nearly 150 n away from the origin of the shaft), the thickness measures not more than 12/><. It then begins to thicken again, up to the maximum E. MARSHALLI. — SPICULATION. 99 thickness of 35 /i. The axial cross has been determined to be (always ?) simple, that is to say, having a single cross-piece. Closely above the position of the cross are the first barb-like spines, of which there are usually two opposite each other, or sometimes more than two in a whorl. The rest of the spines are arranged along the shaft in an approximately regular spiral row. The bundles of basal spicules, so long as they run alongside the skeletal tube, are as a rule entirely covered by the flake- tissues — not exposed by the falling off of the latter. Superiorly they extend in the wall for fully two-thirds of the body-length. The dermalia (PI. IV, figs. 16, 28 ; PI. V, fig. 36)— sword-like hexactins as usual — have rays which are on an average 10 o- thick near the center and taper out to more or less sharply pointed ends. In certidn specimens, however, the hilt-ray was often bluntly rounded at the end, but never showing a swelling in its course. The length of the hilt-ray is 130-170//, on an average 150 (x ; blade-ray generally more than 3 times as long, up to 700/^; guard-rays somewhat longer than the hilt-ray. These form the well-known dermal latticework, the meshes of which are about 200/^- in width. Both the hilt and the blade rays possess for the greater part of their length sparingly distributed, minute tubercles. The guard-rays are nearly quite smooth all over. A specially large size is attained by those dermalia which are situated along the free edge of the cuff and of certain parietal crests and whose hilt- ray stands in association with the prostal oxydiactins already described. These dermalia are frequently nearly 2 mm. long, the hilt-ray measuring 300-400/^ in length. The rays may be twice as thick as in the ordinary dermalia of the general surface, 100 I. IJIMA : HEXACTINELLIDA. I. The gastralia — slender-rayed peiitactins of medium size — are quite like the ordinary dermalia. The paratangential rays, 140-200;« long, are often bent and of unequal length in the same spicule. The distad directed, unpaired ray is straight and frequently more than thrice as long. The gastralia are rather isolated in their distribution, especially so on the inner surface of the excurrent canals, in which they receive the name of canalaria (PL IV, fig. 28). The ßoricomes (PI. IV, fig. 10) are very common, but less so in the elevated regions of the parietal ledges than in more depressed parts. They occur both subdermally and at the apex of dermal hilt-rays. Diameter 70-80 /j. ; principal rays under 6 /^ in length as measured from the central point of the rosette. The number of terminals in a perianth varies from 9 to 12. Marginal teeth of the terminal plate are well developed ; 3-5 in number. The oxyhexasters (PI. IV, fig. 17) are much less numerous than the floricome. They occur both subdermally and subgastrally. Occasionally I have met with some lying outside the layer of dermal paratangentials. They are least numerous, even rare, in the ledges. At certain other places, as, e.g., near the parietal oscula or in the subgastral region, they are tolerably common. Diameter 75-83 /^ Each short and thick principal ray bears 3 or 4, sometimes only 2, diverging terminals. The latter moderately strong, smooth and nearly straight or only slightly bent. The graphiocomes (PL IV, fig. 19 ; see also fig. 28) are probably the commonest of all the hexasters. They occur in all E. MARSHALLT.—SPICULATION. 101 parts of the sponge but are confined to the external trabecular region under tlie layer of dermal paratangentials. The principal rays (fig. 20) are about 7Y2 ,« long as measured from the spicular center ; the surface shows minute tubercles ; the disc at their distal end is, on the outer surface, densely beset with the thickened bases of rhaphidial terminals. When the latter fall ofiP, as they seem to do by a normal physiological process, the bases remain to the disc as small spiny processes (fig. 20). It is of very frequent occurrence that one meets with the relics of graphiocomes after the complete or partial loss of the terminals or rhaphides. These attain a length of 115 fx when fully developed. Free rhaphides are found among the trabeculse in considerable quantities, either irregularly scattered or still grouped in sheaves and often under such circumstances of relative position to graphio- come-relics as put the original connection of both beyond the reach of doubt. They seem to be moved on towards the surface, becoming on the way so directed as to point outwards with one end, and finally to mostly arrange themselves in a bundle-like manner along and around each hilt-ray of the dermalia (PI. V, fig. 36). In no other species that I have studied were the rhaphides so constant or abundant in the position just mentioned. The com- monness, even in old individuals, of the source of rhaphides, i.e., the graphiocome, indicates that these fine needle-like spicules are being constantly thrOAvn out from the sponge surface, in all prob- ability as a sort of defensive missile. The account of develo^Dmental facts with regard to the above three kinds hexasters, I will defer until I shall have completed the histology of the soft parts. As the fourth kind of hexasters j)resent in E. marshalli, I 102 I. IJIMA : HEXACTINELLIDA. I. should here mention the rosette I have figured on PI. IV, fig. 21. It occurs quite rarely and solitarily but apparently is constant alike in both young and old specimens. So far as I know, it seems to be peculiar to the present species. The diameter meas- ures 40-45 ,«. The principal rays resemble those of graphio- comes ; the discs at their ends are lens-like or prominently con- vex on the outer surface. The terminals are exceedingly fine, about as long as or somewhat longer than the principals ; they are pointed at the outer end and arise closely together from all parts of the outer disc-surfiice. The peripherally situated terminals in each tuft are slightly but distinctly flaring, so that the tuft may be said to be campanulate. The entire rosette looks not unlike a plumicome or graphiocome in an early stage of its development in which the terminals are still very short. However, it differs from the former in that the terminals in a tuft are all nearly equally long ; and from the latter it differs in having the termi- nal tuft expanded into a bell-like form. The rosette above described is found only at such long intervals that it requires a close study of preparations in order to come across one. The spiculation of the sieve-plate deserves special notice in respect of a few points. Unlike E. imperialls and many other species, the predominant elements of the sieve-plate parenchymalia are thetactins, instead of diactius. The tlietactins furnish both the principalia and the accessoria, the latter also containing elongate hexactins, pentactins, &c. The parenchymalia, in forming the beams, are disposed in several loose or compact strands, between which are left sufficient spaces for the location of the much folded chamber-layer and of small excurrent canals opening on the i;astral side. All of the four kinds of rosettes found in E. MARSHALLI. YOUNG SPECIMENS. 103 the lateral wall are present in about the same numerical propor- tion. Both the dermalia and the gastralia are almost exactly like those of the lateral wall. The bottom-plate is spiculated in essentially the same manner as the lateral wall, except that here the parenchymalia, which are again mostly thetactins with slender rays, run almost always singly and in various directions, combining but seldom into loose fascicles. The plate is therefore weakly supported and easily breaks down. The dermalia and the gastralia differ in no way from those of the lateral wall. The former are irregularly arranged instead of forming a regularly meshed latticework. Young Specimens. I have before called attention to points by which the early postembryonal stages of E. imperialis and E. marshalli, though closely alike in general outward appearance, may be distinguished (p. 79). Now, from Dôketsba I have not a small number of young Eupledella in various stages of growth, all of which I do not hesitate to refer to the present species. Apart from the circumstances that no other species of Euplectella is known to occur in that locality and tliat the specimens in question form an uninterrupted series leading up to such as have the charac- teristics of E. marshalli fully and unmistakably developed, I have sought to establish the correctness of my identification by ascer- taining the size of the floricome in each specimen. This, in all cases with which we are now concerned, has been found not to exceed 80 /^ in diameter, exactly as it should not if the specimens were E. marshalli ; whereas, in all those young specimens of similar size, which I have referred to E. imperialis, the diameter reached 91 fx. 104 I. IJIMA : HEXACTINELLIDA. I. Ill all the young specimens the tubular body has in general a spindle-like ventricose shape (PI. IV, %s. 6-9). At first the ventricosity is situated at about the middle of the body ; it may however soon be brought to a somewhat lower level, probably because the growth in length at the upper end is relatively more rapid than the general increase in circumference. The lower end is contracted and blindly closed at its juncture with the yet weakly developed basal tuft. The upper end is truncated unless the more or less convexly arched sievs-plate, which is \erj delicate and therefore easily detached in the early stage of its formation, is preserved intact. The smallest specimen I have had was only 18 mm. long (exclusive of the basal tuft), 7 mm. broad at the middle and 2 mm. wide across the round opening at the superior end. Three other very small specimens measured 20-23 mm. in length, 7-9 mm. in greatest breadth and 272-4 mm. across the superior truncated end. One of these specimens is shown in natural size on PI. IV, fig. 6. The entire external surface is on the whole even and covered over uninterruptedly by the dermal layer. A few isolated gaps seen in the superficial tissue proved on close observation to be rents due to laceration. Unless the wall has become untransparent by drying up, the longitudinal beams of the skeleton can be made out to a greater or less extent, while the transverse beams are quite indistinct. The canals in the wall appear to the naked eye as darkish spots, which become smaller toward either end of the body and finally are hardly visible, so that the wall tissue at the ends assumes a uniformly whitish appearance. Of the delicate sieve-plate, which must have been present at the open upper end but which seems to have been lost, I shall speak further on. E. MARSHALLI. — YOUNG SPECIMENS. 105 A very important feature in all the little specimens under consideration is the total absence of parietal oscula. These are at any rate macroscopically still unopened. However, it can easily be demonstrated under the microscope that there exist in the wall, especially at the bulged middle portion of the body, certain canals or niches which, like the typical excurrent canals, extend with a free lumen from the gastral cavity, but, which unlike them, are not blindly capped by the chamber-layer at the outer end. There is, at the spots in question, a gap in the chamber-layer and here the gastral cavity stands in free communication with the outer world through the narrow lacunae between the trabeculœ. At the spots a condition obtains quite similar to that figured by F. E. Schulze in the Challenger- Report ('87, PL LIII, fig. 5) for young Lanvginella j^ujm nt the part where the oscular area should later develop itself. I suppose that functionally the spots in the above described condition are already playing the rôle of parietal oscula. The latter will become definitely established, if only the trabeculse and the dermal skeleton give way for a fieer passage than before, in order to meet the requirement of an ever increasing quantity of outflowing water. I find the above idea concerning the formation of parietal oscula perfectly corroborated by my observations on the growing parts of larger specimens. The character of the wall in quite young Eupleciella might then be said to agree essentially with that of Holascus (see p. 39). The only important difference between the two genera mentioned seems to consist in the development or non-development of parietal oscula. The well-developed external ledges of Eupledella, as also of Regadrella, are apparently the outcome of the excessive 106 I. IJIMA : HEXACTINELLIDA. I. thickening of the wall in the presence of interruptions in the form of parietal oscula. In somewhat older specimens of 32-3Ô mm. length and 10-11 mm. greatest breadth, there already exist a varying number of parietal oscula, which are still very small but yet visible as distinct openings. They first break through in the bulged middle portion of the body. Nearer the ends but especially in the upper region, they are still in a state of incipient formation, being covered over by the dermal layer. The wall-tissue between the open parietal oscula has begun to swell out gently as the first step in the formation of the ledges, which henceforth become more and more conspicuous as the sponge advances in growth. Fig. 7, PI. lY, shows the skeleton of an individual of the size in question, from which the loose tissues have been rubbed off. As representatives of still larger young specimens, whose appearance has notably approached that of adults, will serve the two shown in figs. 8 and 9, PI. IV. In the smaller specimen (fig. 8) the parietal oscula, though yet small in size, are already numerously present. They may be said to be situated at the bottom of shallow dimple-like depressions of the external surface. In many of the depressions, however, the oscula have not yet opened through. In the larger specimen (fig. 9), the externaV elevations between the parietal oscula have definitely taken the form of an irregular network of ridges or ledges, which are most pronounced on the broadest middle portion of the body. Each depressed mesh generally contains only a single oscular opening, but some- times there are more in an indefinite arrangement. The largest of the openings are still under 1 mm. in diameter. They occur E. MARSHALLI.— YOUKG SPECIMENS. 107 Oil the greater part of the body including the lower end. Toward the upper end, they gradually become smaller, while at the same time the external depressions closed at the bottom become more and more frequent. In some of the depressions the oscula are found in the first stage of breaking through. The ledges become superiorly less and less prominent, until finally at a short dis- tance before reaching the superior edge of the lateral wall they cease altogether to exist. So that, there remains at this terminal region an even-surfaced, unperforated zone of nearly uniformly compact appearance, — a zone retaining the characteristics of the wall in a much earlier developmental stage. Such a plain-looking marginal zone is observable up to a stage when the sponge measures about 70 or 80 mm. in length. So long as it persists and also for some time after it has become perforated by newly formed parietal oscula, the soft cuff proper is slightly or not at all developed. This develops distinctly after the ledge formation, which follows that of the parietal oscula, has extended to the u23permost rim of the lateral wall. Never-- theless, I find this rim in all young specimens before they acquire the true cuff not quite thin and sharp, but possessing a firm narrow edge squarely cut off (seethe upper end of fig. 7). This is due to the fact that many of the slender parenchymal spicules composing the longitudinal skeletal beams and coming up to the rim are exceedingly elongated sword-like hexactins, of which the five relatively very short rays, corresponding to the hilt and the guard, are situated in a row at the very edge of the lateral wall. Strange to say, I have not succeeded in finding the same parenchymal hexactins in full-sized specimens. This may how- ever be explained by assuming that the said hexactins after a certain period neither grow in size nor increase in number, anc]. 108 I. IJIMA : HEXACTINELLIDA. I. consequently become so concealed among the crowd of other spicules more lately developed that they easily elude detection. Further it may be that, as the sponge grows in length at the superior end, the parenchymalia once occupying the edge may be left behind, instead of being shifted along .and in perpetual connection with the edge. As to the sieve-plate, in most specimens under oO mm. length, I find it entirely or almost entirely lost, so that the gastral cavity opens above by a wide circular aperture. The beams of the plate in such small individuals are so thin, soft and excessively frail as to break off on the slightest provocation. I have known them to succumb to the rush of water as the freshly caught specimens were being picked out of the sea. At other times I have seen the air-bubble in the gastral cavity or the motion of alcohol into which fresh specimens were thrown, disturb or destroy the sieve- plate. In many cases, the soft and delicate beams were severed clean off from the comparatively firm rim of the wall, leaving no trace of the sieve-plate visible to the naked eye ; in some other cases, they left behind as relics a greater or less amount of shreds attached to the rim. In the specimens figured in figs. 6-8, PI. IV, the sieve-plate Avas entirely gone. Only in two specimens out of several measuring under 50 mm. in length, do I find the plate nearly completely preserved by some fortunate circumstance. I have given double-sized sketches of both these cases in fig. lî> a & b. In the one specimen (39 mm. long), the plate is scarcely or but slightly arched ; the meshes are ten in number. In the other specimen (48 mm. long), it is convex, like a watch-glass ; the number of meshes exceeds 10 b}^ a few. E. MARSIIALLI. — YOUNG SPECIMENS. 109 The beams are at places about Y3 mm. broad, but for the most part are exceedingly fiue. Specimens of over 60 mm. in length of body show a well arched sieve-plate, which is of sufficient strength to remain intact in most cases. For instance, the specimen of fig. 9 (63 mm. long) has a complete, vault-like sieve-plate with about 35 meshes. As to spiculation of young specimens, I have, in the first place, subjected a portion of the youngest specimen I have had (18 mm. long) to a careful examination. The dermalia were found to be sword-like hexactins like those of adults. I mention this because in Regadrella okinoseana I have found the dermalia in quite early postembryonal stages to consist of pentactins which are later replaced by hexactins (see under that species). In the absence of parietal oscula, the oscularia are certainly not devel- oped. It is noteworthy that, although both floricomes and graphiocomes were common and even that rare form of hexasters shown in fig. 21, PI. IV, was met with in a few instances, yet I failed to discover a single oxyhexaster in that little specimen. The floricome (62-72 ij. dia.) was on the average smaller and its terminals somewhat more slender than in adult individuals. Rhaphides detached from the graphiocome were already present in the superficial region, though by no means yet in great pro- fusion. Of the spiculation in other young specimens I will make only the following remark. The oxyhexaster was sought in vain or was met with exceedingly rarely in preparations made from several individuals under 60 mm. in length of body. From about the period of the body-length just mentioned and onward, the oxyhexaster begins to be constantly seen, though not in abun- 110 I. IJIMA : HEXACTINELLIDA. I. dance at first. It occurs, for instance, in the specimen of fig. 9, in some numbers. Tiie oscularia also beoin at about the same time to difierentiate around the oldest formed parietal oscula. Most of them are at first scarcely distinguishable from the gastralia. Essentially the same spicular elements as in adults occur in the delicate sieve-plate, though much more sparingly and in looser arrangement. Concerning the separate beams of the skeletal tube, I may mention that throughout the body of quite young specimens, as in the growing upper end of older specimens, the oblique systems are wanting or at most are represented by isolated fibers. Of the two other systems, the beams of the longitudinal are always somewhat thicker than those of the transverse (fig. 7). This is at any rate partially due to the presence of basal needles in apposition with the former. In order to obtain insight into the mode of development of the skeletal tube, I have counted the number of the transverse and longitudinal beams in a series of not only variously sized young specimens but also of those which might well be consi- dered to be nearly or quite full-grown. The list is given on the following page ; in it the specimens are arranged in the as- cending order of their body-length, beginning with the shortest. I must say that, notwithstanding my efforts to be as exact as possible in counting, the figures in the columns I and II of the annexed table, can represent only approximately the numbers of the beams they stand for. Absolute precision in this matter can not be expected owing to the frequent occurrence, especially in the longitudinal system, of beams of such a character as makes it impossible to decide whether the}»" are to be reckoned in any case as single beams or not. This ambiguity evidently stands in E. MARSHALLT. — YOUNG SPECIMENS. Ill Number of transverse and longitudinal skeletal beams in variously sized E. marslialli. No. Size of Specimens. I. Number of transverse beams. IL Number of longitudinal beams (at middle of body). Length of bodj' (excl. of sieve-plate and basal tuft). Mean diameter of gastral cavity at middle of body. 1 mm. 20 mm. 6 ? 26 2 22 5.5 35 27 3 23 7 30 + 28 4 31 8 39 25 5 32 7 35 27 6 33 7 32 26 7 34 8 32 26 8 34 8 38 33 9 38 7 41 25 10 42 6 36 ? 11 52 10 47 32 12 58 12 39 35 13 QQ 16 62 28 14 74 14 44 29 15 75 11 70 33 16 82 18 43 26 17 88 20 52 38 18 109 20 50 30 19 110 27 44 37 20 116 33 38 43 21 116 38 36 38 22 120 26 36 37 23 122 32 ? 43 24 127 52 28 35 25 130 36 38 43 26 140 30 45 44 27 140 36 51 34 28 143 36 47 39 29 155 40 42 45 30 182 44 40 44 112 T. IJIMA : HEXACTINELLTDA. I. relation to their mode of multiplying themselves, which is by- splitting and gradual separation into two of an originally single beam, as was first shown by F. E. Schulze ('95, p. 24). More- over, the continually recurring rise and fall in the value of fio;ures in the two columns make it manifest that the number of beams in both systems, but especially in the transverse, is sub- ject to certain, often very considerable, variations according to individuals. Under such circumstances there is but little prospect of accurate inductive inferences being made from the annexed table, unless it be known what allowances to make for variations in individual cases, which is certainly not known. Nevertheless, it seems to me that the general trend of figures in columns I and II is anticipated by certain facts observed in the manner of the arrangement of the beams. Firstly as regards the transverse system, it may be considered as a general rule that the distance between each two beams is widest where the sponge-body shows greatest ventricosity, i. e., at or somewhat below the middle of its length. Toward the base, the interval usually lessens somewhat or may remain nearly the same. So also toward the upper end in old specimens. In young specimens, on the other liand, it grows superiorly on the whole gradually and continually less and less, until, at the marginal zone close to the juncture with the sieve-plate, a con- dition is reached which is strikingly different according as the new-formation of transverse beams is taking place or not. It is in that zone of comparatively small specimens only, — say of a length under 70 mm. or 80 mm., at any rate of not over 100 mm., — that there exist indications of the transverse beams undergoing active multiplication. I have before called attention to the primitive character of this outwardly smooth marginal zone, which is yet E. MAESHALLT. — YOUNG SPECIMENS. 113 imperforated by the parietal oscula (p. 107). Here a number of fine circular bands, composed of yet weakly developed parenchy- mal spicules, lie most closely together, attesting their recent formation as was pointed ont by F. E. Schulze ('95, p. 25). On tlie other hîind, specimens of over, say, 100 or 120 mm. length, no longer show tliis peculiar characteristic in the corresponding roo-ion, althontrh this mav still be somewhat backward in the general development of its parts. The last transverse beams at this end of the sponge stand more or less distinctly apart, or at any rate never so close together as in an earlier period of growth, and the very last beam is commonly separated from the superior ritn of the lateral wall by a space which is traversed only by oblique or longitudinal beams that directly pass above into the sieve-plate beams. Parietal oscula are now met with right up to the border of the sieve-j^late. (See upper part of fig. 4, PL IV). To all appearances, then, the marginal region, as also the parts further below, is no longer giving rise to new transverse beams, although the possibility of, so to speak, sporadic new-formations can not be altogether excluded. To sum up : the transverse beams develop to their maximum number before the sponge-body has grown to a length of about 100 mm. During its subsequent growth, the beams should go only wider and wider apart from one another, their number remaining practically stationary or nearly so. Turnius; now to column I of the table, the above fact seems foreshadowed in that some of the highest figures are already met witli before the specimens attiin a length of lOvO mm. The fall of figures for larger specimens, observable in the table, is prob- ably to be explained as mainly due to individual circumstances and not to actual decrease in number. However, I am inclined 114 I. IJIMA : HEXACTINELLTDA. I. to think that the hitter process may at times possihly take place, some of the originally transverse beams becoming transposed into the oblique. Secondly with respect to the longitudinal system (column II), there is observable a general rise in the value of figures from top to bottom of the column, indicating a continual addition to its beams along with the growth of the sponge. This corres- ponds with the fact that even in the oldest specimens there exist here and there such longitudinal beams as seem to represent different stages of splitting lengthwise and separating into two. If now the transverse beams should cease to multiply at an early period while the longitudinal persist in multiplying, the numerical proportion of both in young specimens must be quite different from that in old specimens. This is likely tlie ex- planation of the fact to be noticed in the table that, while down to the specimen of 110 mm. length (No. 19) the excess of dif- ference between the numbers of the two kinds of beams in each specimen stands on the side of the transverse system (I), the case is reversed in most of the remaining larger specimens. For E. simplex, F. E. Schulze ('95, p. 23), by counting and comparing the number of the two kinds of skeletal beams in half a dozen young specimens of various size, has reached the conclusion that, during the growth of tlie sponge, the transverse beams increase considerably in number while the longitudinal do so to but an insignificant degree. To wit : the smallest specimen, 0Ö mm. long, had 25 transverse and 28 longitudinal beams against 40 and 30 respectively of a specimen 110 mm. long. In order to see how far Schulze's above conclusion can be verified with E. mai'shalU, it would be necessary to take into considera- tion only those young specimens of that species in wiiich both E. MARSHALLI. — YOUNG SPECIMENS. 115 systems are actively increasing the nnmber of their beams. As such may be regarded Nos. 1-17, given in the table (say, speci- mens ranging from 20 to 90 mm. in length). From the data afforded, I think it safe to give the range of numerical increase, in these S23ecimens, of the transverse beams as from 32 to 52 or even up to great deal more (say, an increase of 20 and over), and of the longitudinal beams as from 25 to 38 (an increase of 13, which is likely about the maximum limit). It can not then be gainsaid that, during the life-period represented by these specimens, the former increase in number with greater rapidity than the latter, — a fact which, so far as it goes, conforms to the general tenor of Schulze's statement. Should however a young specimen of E. marshalli be brought into direct comparison with an old one as regards the points in question, one may be misled into quite different inferences. For instance, by comparing specimen No. 2 (of the table) with No. 28, the appearance is that the rates of increase of the two kinds of beams have kept pace together, both showing alike an increase of 12. It may even be found, as e.g. by comparing Nos. 2 and 30, that the longitudinal beams have increased far more than the transverse. It is plain that these appearances are due to the fact that the longitudinal beams have continued to multiply themselves after the transverse have ceased to do so. I think what has been said above concerning the increase of the skeletal beams is, in the main, applicable to all species, or at least to those in which the synapticular fusion of spicules never takes place. In E. imperlalis, which is one of the species with a partially rigid skeletal framework, the multiplication of the beams is made impossible as soon as the amalgamation of their elements sets in and so far as this extends in the lower 116 I. IJIMA : HEXACTINELLIDA. I. part of the body. At the upper end, however, it still continues to go on (cfr. pp. G8, 82). Here, the new-formation of the trans- verse beams is carried on apparently to a much later life-period (as judged by the size) than in E. marshalli, but it likewise seems to stop some time before that of the longitudinal beams is brought to a completion or this end of the body attains its maximum girth. Soft Paets. As the most readilv obtainable Hexactinellid in the Saoami Sea, E. marshalli has supplied me with my best opportunities for the study of the soft parts. The following account, in the absence of special mention to the contrary, refers to that specie?, although in the main it may be regarded as applicable to a wider circle of forms and even to the Hexactinellida in general. As has been stated by F. E. Schulze ('87, p. 23), the histological structure is so uniform throughout the entire group, that the modifications to be noted are hardly of an important character. Let it at once be stated that as regards the general arrange- ment of the soft parts, the facts before known through the in- vestigations of F. E. Schulze ('80, '87, 'gg«, ig'a), so far as they go, have found essential confirmation in the results of my own observations, though in respect of certain important points relating to the finer structure, my results stand irreconcilably at variance with his. The sponge-wall, being composed of the soft parts and the spicules, is, as has been observed by F. E. Schulze, remarkable for the exceedingly cavernous character of its structure. It is thoroughly penetrated by intercommunicating lacunar cavities and passages, across which, it may be said, the soft parts stretch E. MAESHALLI.— GEN. AERANGEMENT OF SOFT PARTS. 117 themselves only in the form either of cobweb-like or film-like trabeculse or of a cribellate membrane (chamber-wall, membrana reticularis). All the soft parts, if put together apart by them- selves, would make up but a comparatively small volume falling considerably below the mass of the spicules and would appear almost insignificant in proportion to the space occupied by the entire sponge in its undisturbed state. General arrangement of the soft parts and their RELATION TO THE SPICULES. — The flagellated chambers, whose structure will be specially dealt with in the next paragraph, are, as is well known, arranged side by side in a single layer, the chamber-layer (PI. IV, fig. 28, ch. I.), which separates the outer from the inner trabecular layer of the sponge-wall (see p. 41). As is further known, the chamber-layer (which is not to be confounded with the chamber-wall) forms in the choanosorae numerous, outwardly directed protuberances or evaginations, which are proximally open and distally blindly closed. The evaginations aie of various sizes and of great complexity of form. While some are quite small and simple, others, especially those that extend into the parietal ledge and correspond in position with the larger excurrent canals, may be of very considerable length and caliber, and moreover bear on their sides a greater or less number of secondary evaginations, which may again repeat the branching process. The final branches belonging to different systems of the evaginations remain separate, although the possi- bilit)'- of their coming into an intercommunicating anastomosis under exceptional circumstances can not be excluded. The chamber-layer may then be considered as forming by itself a voluminous mass with its complex system of evaginated 118 I. IJIMA : HEXACTINELLIDA. I. protuberances. Assuming that mass to be free of the trabeculae and the spicules, with which it in reality is connected both externally and internally in making up the choanosome, there should lead out proxiraad from it numerous, separately opening tubular passages, not unlike the radial tubes of Sycons and which give molding to the excurrent canals of the choanosome. Ex- ternally, between and around the contiguous, irregularly shaped protuberances, there should exist, this time not tubular passages, but an exceedingly intricate and continuous interspace, which, like the intercanals of Sycons, forms a part of the general system of incurrent spaces. iVlong the course of the incurrent canals penetrating into the choanosome, the interspace just mentioned is more or less wide ; at most other places, especially in the deeper region, it is quite narrow and often exceedingly com- pressed in that the external surfaces of chambers, belonging either to the same or different protuberances, come nearly or even quite into contact with one another. The trabeculse in their relation to tlie above mass of the chamber-layer show in the original primitive condition the following arrangement : The outer trabecular system forms a continuous superficial layer, coveiing over the outer ends of the chamber-layer evaginations and thence extends alike into all parts of the afferent interspace between the latter ; similarly, the inner trabecular system continuously covers the inner surface of the sponge-wall and also pervades all the efferent hollow^s of the evaginated chamber-layer. In shoit, we may consider the entire thickness of the sponge-wall as consisting of a nearly uniform network of thin trabeculœ which keep the folded chamber-layer suspended midway between the two surfaces of the wall. Such a comparatively simple arrangement of the soft parts is alw^ays E. MAESHALLT. — GEX. AKRANGEMENÏ OF SOFT PARTS. 119 met with in that region of the body, especially in very young specimens, in which the wall is still thin and backward in the development of its parts. As is easily conceivable, the increase of the wall in thick- ness and of the chamber-layer in the extent and complexity of its evaginations, puts into requisition a freer passage than before for the accelerated iu-flow and out-flow of water, and thus arise the iucurrent and the excurrent canals. Both of these systems of canals are simply relatively larger intertrabecular spaces which, in the form of elongated passages, penetrate more or less deeply into the choanosome. The canals are therefore, at the commence- ment of their formation, indistinguishable from ordinary inter- trabecular lacunae. However, after attaining a certain length and caliber, they deserve their name all the more since the lining trabeculse and certain spicules give to them a more or less definite, though of course much interrupted, septum-like wall. The excurrent canals (PI. IV, fig. 28, ex. c.) develop each as a direct continuation of the gastral cavity in the axis of the efferent hollows inclosed in the evaginations, before mentioned, of the chamber-layer. They therefore not only correspond in their position with, but also repeat to a great extent the branched configuration, of the latter. The result is that the canals directly communicate with the gastral cavity by widely open orifices, which, unlike those in many species belonging to other families, are not covered over by a continuous endosomal layer supported by a lattice-work of gastralia. The internal trabecular system, forming a thin layer, is directly continued from the gastral surface into the evaginations of the chamber-layer, along the inner surface of these and around the lumen of the excurrent canals. Toward the ultimate branches of the evaginations and 120 I. IJIMA : HEXACTINELLIDA. I. after tliese have diminished in caliber beyond a certain limit, the canalar lumen disappears or rather becomes indistingaishable from the ordinary intertrabecular lacunse ; in these small branches, as in fact in all evaginations of insignificant dimensions, there persists a primitive condition in that the entire internal space is traversed uninterruptedly by the trabeculse. The incurrent canals, it scarcely needs to be specially pointed out, are canalar gaps in the outer trabecular system which per- vades the external recesses between the evaginated protuberances of the chamber-layer. They branch during their inward course and may undergo anastomosis with their fellows. They are not always circular in cross -section. In all Hexactinellida they are as a rule smaller but more numerous than the excurrent canals and further unlike these, they never break through externally so as to open directly onto the surface of the sponge-wall. With their outer apertures they join the lacunar spaces in the peri- pheral trabecular layer, and in fact all the lacunse and cavities among the trabeculœ form one intercommunicating system on either side of the chamber-layer. In the peripheral or superficial layer of the external trabe- cular system just referred to, and which I have before mentioned as continuously covering over the outer ends of the chamber- layer protuberances, there may be distinguished two strata, the outer ectosomal, and the inner subdcrmal stratum ; although it must not be imagined that there always exists any sort of a well-defined demarcation between them. The ectosomal stratum or the eclosome is the seat of the latticework of the dermal skeleton and is more or less speciulized in consequence of that fact as well as of its most superficial situation. The subdormal stratum is characterized by its relatively more spacious lacuna) E. MAESHALLI. — GEN. ARRANGEMENT OF SOFT PARTS. 121 or, wluit amounts to the same thing, by the comparative sparse- ness of trabecuUxî. The hicunse in this region are known as the subdermal cavity (PI. V, fig. 36, s.o.). It is from this cavity that the incurrent canals appear to arise, thence to pene- trate into the choanosome. Thus, generally speaking, the ectosome extends itself over, and is separated from the choanosome by, the subdermal cavity ; the paratangential rays of the dermalia serve as its main support, while the proximal rays of the same as well as a variable quantity of subdermal trabeculœ, effect its connec- tion with the choanosome. As it presents itself in Euplectellidse, a considerable thick- ness is to be ascribed to the ectosome, a fact which is apparently caused by the presence of well-developed distal rays to all the dermalia. Each of the rays just mentioned stretches out the thickness of the ectosome in distal direction so as to form a minute conulus on the external surface of the sponge. The boundary delimiting the ectosome from the underlying subdermal stratum is about as ill-defined as can be. This is in a great measure due to the fact that the subdermal cavity never reaches a prominent degree of development in spaciousness, a peculiarity which stands in correlation both with the fact that numerous proximal rays of the dermalia traverse the region at comparatively short intervals, and also with the small caliber of the incurrent canals. Nevertheless, I think there are grounds for considering that the plane of the dermal latticework (paratangential rays of the dermalia), which in Euplectellidse lies, as is well known, at a certain distance below the external surface, indicates in a general way the boundary between the two strata. The ectosome is there- fore not to be described as a perforated plate-like layer, but, as a part of the general trabecular system, it consists, throughout its 122 T. IJIMA : HEXACTINELLIDA. I. entire thickness, of trabeculse in an irregular cobweb-like arrange- ment. On the whole the trabecular cobweb of the ectosome is somewhat denser than in the region below the dermal paratangen- tials, i. e., in the subdermal region (see PL V, fig. 36). It is important to notice that in the trabecular cobweb of the Euplectellid ectosome in general, the most superficially situated trabecule, i. e., those delimiting the sponge periphery from the external world, are often, but not invariably, expanded paratan- gentially in a film-like or membrane-like manner. The gaps, or the ' pores,' inclosed by such flattened trabeculœ are of a more or less roundish shape and give to the layer itself the appearance of a perforated membrane. This has been called by F. E. Schulze the ' dermal membrane,'' and accordingly, the dermalia, as being situated beneath that membrane, have received the name of ' hypodermalia.' Misleading and inappropriate as the latter appellation seems to me to be (see p. 46), the former may with advantage be retained for the purpose of description. The dermal membrane then forms only a small part of what I have called the ectosome in Euplectellidœ. Tn other families in which the distal rays of the dermalia do not come into development, the ectosome becomes, as suggested on p. 46, greatly reduced in thickness in that the dermal membrane is brought down to the level of the dermal paratangentials. It is all the thinner because of the subdermal cavity which is generally more spaciously developed in those forms than in Euplectellidae. The dermal membrane then stands nearly or quite by itself for the soft part of the entire ectosome, in which case the two names may be considered as practically synonymous. The ectosome of E. marshalH in particular requires a few E. MARSHALLI. — GEN. ARRANGEMENT OF SOFT PARTS. 123 more words of comment. Noteworthy is the fact that in that species the dermal membrane is scarcely sufficiently developed to deserve being called membranous. In other words, the limiting trabeculœ of the external surface are generally as thin and cobweb-like as, and in no way distinguishable from, those of the deeper parts. The meshes of the surface, or the ' pores,' bounded by such trabeculœ are exceedingly various and irregular in shape and size, just like any intertrabecular lacunœ seen on sections of the sponge-wall. However, occasionally in the spaces between the conuli the limiting trabecular are found flattened out into the form of a narrow band or of a nodal expansion (PI. IV, fig. 23), which, so far as it extends, gives a more or less rounded outline to the meshes bounded by it. In E. aspergillum, as described and figured by F. E. Schulze, the dermal membrane should be well developed as such ; so I have found it likewise in E. imperialis, or, at any rate, more extensively membranously formed than in E. marshalli. On account of the cobweb-like nature of the entire ectosomal trabeculai in the last mentioned species, the conuli, when seen from the sides, appear more like the rigging of a schooner's mast (PI. V, fig. 36) than like conical tents, which they would certainly resemble if only a continuously developed dermal membrane were present.'-' * W. Marshall ('75, fig. 62) has described and figured the dermal membrane of a young E. aspergillum as regularly possessing a single, rather small pore to each quadrate mesh of the dermal latticework. This has been shown by F. E. Schulze ('80, p- 666) not to hold true in old specimens, in which the pores had been found to be more numerous and crowded so as to leave less space between them. However, Schulze declared himself willing to believe that in the young the pores might be distributed in the manner described by Marshall. In quite young specimens of E. imperialis as well as of E. marshalli, I find not only the dermal membrane but also the gastral and the canalar membrane respresented by quite thin trabecuke, which are nowhere membranously developed. I regard this as the primitive condition of the trabeculœ at the surfaces and the membranous state as being acquired after a certain stage of growth. 124 I. IJIMA : HEXACTINELLIDA. I. In connection with the ectosome let me here say a word about the endosome. I consider this as being in its fullest development in those Hexactinellida, as, e.g., most Rossellids, in which the gastral skeleton is so highly developed as to form a continuous latticework covering the inner apertures of the ex- current canals. In such cases, the main spicules of the lattice- work— the autogastralia — are generally hexactins disposed in much the same way as the hexactin-dermalia of the Euplectellid ectosome ; consequently, the endosomal trabeculse connected with the autogastralia likewise exhibits an arrangement more or less similar to that in the Euplectellid ectosome. Now, in Euplectella, as also in certain other genera, the gastralia are far too few to form a continuous latticework; so that the inner apertures of the excurrent canals remain perfectly open. Moreover, the gastralia present are pentactins lacking the freely projecting proximal ray ; and what here exists of the endosome between the excurrent apertures, is represented merely by the trabeculae delimiting the internal trabecular layer from the gastral cavity, very much in the same way — mutatis mutandis — as the ectosome (dermal mem- brane) is represented in those species in which the freely out- standing distal ray is wanting to the dermalia. The said trabeculsG (not excluding those of E. marshalli), though often cobweb-like and indistinguishable from those more deeply situated, are at places spread out into a more or less extensive film-like membrane, a circumstance which makes the name gastral membrane appear all the more applicable to them inasmuch as they make up the lining of the gastral surface. The gastral membrane is continued into the excurrent canals as the canalar membrane. The intertrabecular lacuna? underlying the gastral and the E. MARSAALLI. — GEN. ARRANGEMENT OF SOFT PARTS. 125 caiialar membrane (subgastral and subcanalar lacunœ) are never specially widened as the subdermal lacunae or cavities are. Hence, the layer occupied by them (i.e., the internal trabecular layer) is much thinner and contains a much less quantity of the trabeculse than does the peripheral trabecular layer outside the choanosome. Haviijo" dealt with the ectosome and the endosome in their relation to the skeletal parts supporting them, it may not be amiss here to complete our account of the relation existing between the soft parts and the spicules in the choanosome. Leaving the hexasters out of question, the spicules and spicular parts that enter into the composition of the choanosome are : 1) the entire parenchymalia, 2) the proximal rays of the dermalia and 3) the distal rays of the gastralia. All these are distributed on either side of, and nearly completely separated into an outer and an inner set by, the chamber-layer. At the oscular edge the chamber-layer ceases to exist and the two sets of course mix together. In other situations I hold it exceedingly doubt- ful if there exist any spicules which penetrate right through the chamber-layer. The point is rather difficult to settle by the examination of sections and still more so by any other method, but I have never once noticed, not only in Euplectella but also in any other Hexactinellid that I have studied, a spicular ray which undoubtedly passed through the chamber-layer. Certain it is that the wall itself of the flagellated chambers is never pierced through by spicules ; so that if ever a spicule does extend across the layer, it must do so between the separate chambers. I should think that the chamber-layer, in extending itself and making evaginations with the growth of the sponge, pushes its way in the intervals between the spicules present and that 126 I. IJIMA : HEXACTINELLIDA. I. subsequently the growth of the spicules on either side of the layer takes place ivithin the limits of their respective trabecular systems. The outer set of the choanosomal spicules exhibits a much greater development than the inner in the number of individual spicules, some of which here attain their largest size and also frequently group themselves into compact fascicles. This peculiar- ity is evidently correlated with the relatively large and continuous extent of, and the abundance of trabeculœ in, the space occupied by the set in question. Included in the set are all the numerous and long proximal rays of the dermalia and the greater part of the entire parenchymalia. Amongst the latter belong here the most important parts of the skeleton, viz., all the beams of the skeletal framework. These lie apparently in the deepest fundus of the recesses belonging to the external trabecular layer and consequently close to the gastral surface of the sponge-wall. We may consider the chamber-layer as properly lying closely inside the framework and as forming protuberances wherever the meshes and other interstices of the latter permit. Exactly the same relation plainly obtains between the dictyonal framework and the chamber- layer in the Dictyonina. In Eupledella it is not always easy to clearly make out the relation on sections, owing to the confusing intermixture of spicules and chambers ; however, the appearance of the numerous small and shallow excurrent canals occurring all over the gastral surface of the skeletal beams sufiiciently attests the presence of the chamber-layer iîiside the beams. Of the rest of the parenchymalia belonging to the outer set, a small portion runs in a thin loose layer on the external choanosomal surface over the blind ends of the chamber-layer evaginations, while a by far larger portion traverses in all directions the interspace E. MAESHALLT. — GEN. ARRANGEMENT OF SOFT PARTS. 127 between the said evaginations in association with the straight penetrating proximal rays of the dermalia. The said spicules and their bundles stand of course in connection with the trabecular system of the region and leave open the external apertures as well as the lumen of the incurrent canals, to which they partially furnish an incomplete wall. On the other hand, the inner set of the choanosomal spicules is comparatively very weakly developed, a fact corresponding to the sparseness of the trabecule and the thinness of their layer on that side. In the first place, the distal rays of the gastralia and of the canalaria are neither so numerous nor so long as the proximal rays of the dermalia on the outside. Further, the parenchymalia of the inner set, which in part go to supplement the gastralia and the canalaria in supporting the respective lining membranes, are decidedly not numerous ; they are moreover all thin and if grouped at all, appear at the most in thin loose strands. All the soft parts alluded to in the above account are found in the cuff, in the sieve-plate beams and in the bottom-plate in essentially the same arrangement as in the lateral wall. I .emphasize this fact, because it clearly manifests the identical nature of the body-parts just mentioned and furthermore serves to give basis for regarding all the large gaps in the lateral wall, in the sieve-plate and in the bottom-plate alike as oscula (see pp. 38, 39, 94). To follow, by way of a resume, the course of w-ater in its passage through the sponge-wall : Through the pores of the dermal membrane and the intertrabecular lacunœ of the ectosome, it enters into the subdermal cavity. Here it almost directly bathes 128 I. IJIMA : HEXACTINELLIDA. I. the external surface of the most peripherally situated flagellated chamhers ; a large portion of it, however, passes, in order to reach the more deeply situated chamhers, into the intertrahecular lacunœ between the chamber-layer evaginations, at places directly and at other places by means of the excurrent canals. As will be shown in the next chapter, the water enters into each flagellated chamber through innumerable minute prosopyles to find exit by a single large apopyle into the intertrahecular lacunœ inside the chamber-layer. It then finds its way out into the gastral cavity either directly through the gastral membrane or by w^ay of the excurrent canals after passing through the canalar membrane, according as the chambers are situated in the deepest part of the choanosome or in more peripheral positions. Final discharge takes places through the oscula in the sieve-plate, in the lateral wall and in the bottom-plate, — probably most ener- getically through those of the first named structure (sieve-plate meshes). I will now proceed to give the results of m}^ observations on the structural details and relations of the flagellated chambers, of the trabeculœ, &c., beginning with the Flagellated chamber. — Tn shape the individual cham-. hers are generally cup-like, thimble-like or glove-finger-like (see PI. IV, fig. 28)." They mostly measure 80-200//- in length, *This, as is well known, is the most usual form of the Hexactinellidan chamber. Quite a different development may however be attained by tlie ohambers of certain Hyalonematid species, tliough in all probahility as the result 'of the secondary branching and anastomosing of the original saccular form. For instance, in Jlyalonema aßine Marsh, and Sericolophus reflcxus (Ij.) {= Hynhinema rejlexum I J.) I have determined after a careful study tiiat the chambers are represented by an intercommunicating system of canals, whose wall consists of the membrana reticularis. Tlie general arrangi'inont of the flagellated canals strongly reminds one of the configuration of a Fan-ca colony. E. MAESHALLI. — FLAGELLATED CHAMBER. 129 occasionally being as long as 11Ô ii. Those situated in the peri- phery of the choanosome and adjoining the subdermal cavity are, on the whole, somewhat longer than others in deeper positions. The cross-section is approximately circular, with a diameter of 45-90 // (about 7'") ','■ on an average). The cavity within the delicate chamber-wall is always empty, — I mean, perfectly free of trabeculfç. The broadly open, truncated end is the so-called apopyle, by which the chambers open into the lacunœ of the internal trabecular layer. On nearly the entire external surface they expose themselves directly to the incurrent lacuna3 of the external trabecular layer. Here and there on that surface the fine branched ends of the external trabeculai find their insertion (PI. V, fig. 36). Sometimes the chamber shows one or more outbulgings on the sides or near tlie outer end, and these may sometimes be so prominently developed as to bring about the appearance of a lobed chamber. In such cases the wall (reticular membrane) passes without doubt continuously from one lobe into another, making a sharp or a rounded bending. I think these outbulgings or diverticula indicate the process by which the chambers multiply themselves. After reaching a certain stage of develop- ment, the daughter-chambers should become histologically dis- continuous and acquire a certain degree of independence, though remaining side by side and connected together in the manner soon to be described. In forming the chamber-layer often alluded to, the fully formed chambers are arranged close together with the apopyles all directed the same way, without however mutually pressing one another at any point. Hence, the chambers as well as the apopyles remain round or roundish iu circumference, leaving an 130 T. TJIMA : HEXACTINELLIDA. I. interspace between and around them except in places along the convex sides of any two adjoining chambers, where, as it not infrequently happens, the outer surfaces of these may lie to a ofreater or less extent in direct contact with each other. The point in question may best be studied on either real or optical sections across the chambers. The said interspace, whicli scarcely needs be pointed out as a part of the in current lacunar system of the external trabecular layer, is broadest where three or some- times more chambers at a time give boundary to it (PI. V, fig. 43). The corners of such a space extend into the exceedingly narrow cleft between every two chambers, which cleft, as above men- tioned, may at times be obliterated by the coming in contact of the opposite surfaces. Even in the latter case, I think, the contact surfaces, so far as the parts of the wall concerned show the characteristic reticular structure, are not in actual fusion, though sometimes a fusion may occur at points where the chambers come in contact at the rim (marginal membrane), in which part the chamber-wall is, as will soon be seen, structurally the same as the trabecul^e. The intercameral space above referred to is traversed by fine branching trabecule (fig. 43, ti\), which extend between chamber- walls or connect these with the supporting spicules and serve to keep the chambers expanded and in position. Close to the chamber-rim the trabeculœ are replaced by an exceedingly thin, fairly continuous membrane — the connecting membrane or mem- brana reuniens of F. E. Schulze (PI. IV, fig. 22, cm.) — which thus spans the interspace left between the circular apopyles and joins these together. The connecting membrane may be said to shut off the intercameral incurrent space from the excurrent lacuna? of the internal trabecular layer. Occasionally there exist E. MAESHALLI. — FLAGELLATED CHAMBER. 131 open gaps in it ; however, since its situation is such that the current of water caused by the flagella in the chambers could scarcely exercise an unequal pressure on it in either direction, there should be practically a standstill of water at the gaps under normal circumstances. On both of its surfaces the membrane furnishes points of insertion for several trabeculse. As to its morphological nature, I believe in its identity with the trabeculse. It seems to consist of trabeculse simply spread out in a film-like manner, just like certain parts of the dermal, the gastral or the canalar membrane. The thin protoplasm looks exactly like that of any local trabecular expansion ; the nuclei, met with at long and irregular intervals, are just the same in size, appearance and staining capability as those of the trabeculse. Moreover, where several gaps lie close together (as on the left- hand side of fig. 22), the thin beams left between them are in no way distinguishable from the ordinary trabeculae. Turning our attention to the chamber-wall itself, this con- sists par excellence of the reticular membrane formed by the choanocytes and of a narrow and fihny rim around the apopyle, which rim I will call the marginal membrane. Let the latter be first treated of here. The marginal membrane (PI. IV, fig. 22 tfe PL V, fig. 39 ; m.m.) is but another structure which is to be considered as identical in nature with the trabeculse. It is in no distinguishable feature different from the connecting membrane. Like this it is occa- sionally fenestrated and where the gaps occur close together, the appearance is exactly like that of a trabecular cobweb. The protoplasm is seen at places to be directly continuous with the trabecular arising from or inserted in it (figs. 22 & 39). The 132 I. IJIMA : HEXACTINELLIDA. I. nuclei (fig. 39, tr.n.) are indistinguisliable from the trabecular nuclei, but quite distinct from the choanocyte nuclei with which they may lie side by side. The edge of the membrane around the aj)opyle may be said to be even and free, except for the isolated trabeculie which may sometimes proceed directly from it. At a short distance from the free edge, the marginal membrane passes into the reticular membrane and at the same time into the connecting membrane as well. With the latter it is uninter- ruptedly continuous. Into the former it merges rapidly but with- out any definable demarcation, as will best be judged from fig. 39. I might as well have described the connecting and the marginal membrane as one and the same part with which the reticular membrane comes into juncture. Of quite usual occurrence is the fact that the chambers are somewhat contracted at the marginal membrane ; so that, when looked at from either the outside or inside of the apopyle at that end, this appears surrounded by a narrow ring of the marginal membrane in a manner that reminds one of the velum in Craspedote Medusae (fig. 22, m.m.). As seen in optical sections passing lengthwise through consecutive chambers, the opposite side-walls of two adjoining chambers run down close together, or perhaps in direct apposition, toward the rims, finally to diverge more or less from each other when they come to the marginal membrane. Shortly before they freely end, the connecting membrane, likewise in optical section, stretches across between them. However, it must not be thought that such a state is invariably found. Sometimes the chamber-wall runs straight out when they come to the marginal membrane, which may then, in certain parts of its circumference, lie in contact and probably in fusion with the marginal membrane of the neighboring chamber. E. MAESHALLI. — FLAGELLATED CHAMBER. 133 A relation similar to that between the reticular and the marginal membrane in the chamber- wall seems to be repeated in the transition of the chamber-layer into the general trabecular system. Such a transition should occur around every osculum, where the chamber-layer must have termination. In Euplectella I have not been able to bring this termination into view, probably in consequence of the complicated folding of the layer close to the oscular edge. Whereas, in certain Hyalonematids and Rossellids it was not difficult to determine on both sections and surface-views that, close to the thin oscular edge, the chamber- layer was represented by an irregularly undulating, continuous sheet of the reticular membrane, whose reticulation finally be- came merged into, and indistinguishable from, that of the trabe- cular system. The reticular membrane, or as it has been called by F. E. Schulze the membrana reticularis (PI. V, figs. 36-43), forms one of the most characteristic features in the organisation of the Hexactinellida. It consists of peculiar choanocytes whose flattened and ramified bodies, in my opinion, join with one another to constitute an extremely thin and delicate layer of minutely meshed network. When seen under a microscope of moderately high power, the reticulation presents an elegant and tolerably regularly checkered pattern. Under a very high power, the pattern loses in appearance much of its regularity. The meshes, though mostly quadrate, are frequently trapezoidal, rhomboid or triangular in shape, with usually rounded corners. The sides measure 3-7 /^ in length. From all that I have seen of the reticular membrane not only in Euplectella but also in a series of Hyalonematid, Kossellid 134 I. IJIMA : HEXACTINELLIDA. I. and Dictyouine species/*' I can not but maintain that the meshes in question are all open and admit of a free passage of water from the incurrent lacunae into the interior of the chamber. Only when the specimen is badly preserved or when the proto- plasm of the choanocytes is insufficiently stained, is difficulty experienced in deciding whether or not the mesh spaces are over- spread and closed by a transparent membrane, but in successful preparations the contour line of the reticular beams stands out sharp and distinct against the perfectly empty meshes, so that there can be no doubt whatever of the freely open nature of each and every mesh in the reticular membrane. That this is not due to the drastic effect of the preserving reagents, I have fully satis- fied myself by repeatedly comparing the results of experiments conducted according to different methods (see p. 34). I regard all the numerous meshes of the reticular membrane as representing so many prosopyles. There exists among them none that is particularly distinguished from the rest by a specially large size or by a rounded shape. The above stands in marked contrast to the condition we usually observe in other sponges. As is w^ell known, the prosopyles in certain forms occur in tolerably large numbers to each chamber, but these always break through the choanocyte epithelium in a scattered distribution. Whereas, in the Hexactinellida they are to be con- sidered as establishing themselves in all available interstices between the individual choanocytes, converting the epithelium into a veritable sieve-membrane. This state, in my opinion, arises, because of the minimum development or, more probably, of the utter non- development, of mesoglœa in the parenchyme *Tlie reticular membrane of raaay of these species will be figured and remarked upon iu future numbers of this series of Contributions. E. MAESHALLT. FLAGELLATED CHAMBER. 135 (' mesoderm '), which fact causes the excessive thinness of the trabecuhc and the direct exposure of nearly the entire external surface of the chambers to the proportionally widened incur- rent lacunœ. I will return to this point again when I come to speak of the trabecule. Here let it be remarked that the poly- prosopylar chambers of certain non-Hexactinellids seem to pre- pare the way for the condition seen in the Hexactinellida, and that an opposite departure in the structural respect under consi- deration is found in those sponges which have diplodal chambers and which are usually remarkably compact and fleshy on account of the voluminous development of the parenchyme. The beams of the reticular membrane are generally flat and narrow bands of variable width. Here and there, they are very thin and thread-like (see figs. 37, 38). The nodal thickenings are formed by the convergence and juncture of the beams, generally four, but sometimes three or five, in number at each node. They usually contain each a single nucleus (rarely two nuclei lying side by side) and therefore represent the central portion of the choanocyte body, of which the reticular beams are but lateral processes. Sometimes in its course the process is seen to give off a branch or branches, which go to unite either with an adjacent beam or a node. When the membrane is looked at en face, the nucleus appears circular and is surrounded by a proto- plasmic area, which is drawn out into the lateral processes. In profile view or optical section, the nucleus presents an elongate elliptical shape, indicating its marked compression on the plane of the reticular membrane. (See figs. 40, 41, 43). The node itself being likewise compressed, there is scarcely visible a protoplasmic layer on either surface of the nucleus, though at the poles of its elongated axis there exists a small protoplasmic accumulation 136 I. iJBrA : hexactinellida. i. which is directly continuous with tlie thin lateral processes. I have noticed no appreciable difference in the convexity of the two sides of the nucleus, nor have I detected the presence at its distal surface of a strongly colored cap-like body, such as was seen by F. E. Schulze in Schaudinnia arctica. The protoplasm of the nodes and beams, as it appears in hardened preparations colored with carmine or hœmatoxylin, both of which usually stain it but very faintly, consists of a clear matrix containing granules that are neither uniform in size nor in distribution. The limit of the matrix against the meshes is often scarcely perceptible, on account of its perfect clearness ; in fact, the protoplasm presents itself to the eye almost by its granules alone. These are, as have also been noted by F. E. Schulze, frequently arranged in strings, an arrangement which seems to me to be simply due to their situation one behind another ia narrow tracts of the matrix. In some preparations and sometimes in certain parts of a single preparation, I have found in greater or less abundance unusually coarse and réfrin- gent granules (fig. 38), wdiich elsewhere are either quite absent or only solitarily present. They remain unstained by borax- carmine but readily take up acid-fuchsin. Their presence or absence presumably depends upon certain metabolic conditions of the choanocytes ; they are possibly somewhat allied to, if not identical with, the inclosures of the thesocytes to be described further on. Stained with acid-fuchsin, the protoplasm becomes tolerably well colored. It then appears at places nearly homogeneous or uniformly finely granular, and in other places with the coarser well-stained granules in addition (figs. 37, 38). As before men- tioned, its external limit against the meshes stands out well E. MARSHALLT. — FLAGELLATED CHAMBER. 137 defined, without however showing the slightest indication of the presence anywhere of a limiting membrane. F. E. Schulze, as he described the structure of the Hexactinellidan chamber- wall for the first time, evidently believed that the choanocytes rested on a continuous basal membrane, the outer (incurrent) surface of which was furthermore assumed to be lined by a pavement- epithelium. This was a mistake, Schulze ('99a, p. 209 ; 19' a, p. 98) himself has been led by his recent researches into the histology of Schaudinnia aretica to the conviction that the mem- brane has no existence, and that the presence of the pavement- epithelium is questionable. I think it may be considered as a settled question that there exists no special layer of any kind outside of, and in contact with, the reticular layer of the choano- cytes. Schulze {I.e.) has expressed the opinion that the fundament on which the latter layer lies, should be considered to be a relatively wide-meshed network of certain trabecule. I should rather say, as I have already said above, that all the trabeculie coming to the reticular membrane simply find inser- tion in this for the ends of their fine dentritic branches. Some of these terminal branches are indeed seen to creep along the outer surface of the reticular membrane shortly before they terminate ; but such occurrences in the species studied by me are decidedly too few and far between to be regarded as giving a * Grundlage ' to the choanocyte layer. The nucleus, whose shape and position have already been described, is plainly visible under a high magnifying power. It measures only 1.5-1.7 /^- in diameter as seen in the surface view of the chamber-wall. A fine nuclear membrane seems to be present. The contents are nearly homogeneous or at most finely 138 I. IJIMA : HEXACTINELLIDA. I. and sparsely granular. As known through F. E. Schulze, the nucleus is remarkably poor in chromatin, on which account, it, unlike all the nuclei of other tissues, does not surpass the surrounding protoplasm in staining capacity. Nor does it con- tain a body which might be called the nucleolus. Hence, in certain Hexactinellid species or in a certain state of preservation, the nuclei of the choanocytes can be demonstrated only with difficulty. In the surface view of the chamber-wall stained with acid-fuchsin (fig. 37), a clear ring is observable around the nu- cleus at a certain focus of the microscope ; this is apparently due to the refrangibility of the nuclear substance. Also a highly refractive spot is visible in a central position in each nucleus ; this is due to the origin of the flagellum, which arises directly from the inner (distal) surface of the nucleus, and should not be mistaken for a nucleolus. To F. E. Schulze belongs the credit of having first demon- strated the Hexactinellidan flagellum and collar in Schaudinnia arctica. The former structure had been long known to me from Eupleciella marshalU and Acanthascus cactus, in both of which it is fairly constantly preserved in preparations fixed with cor- rosive sublimate. As the total length of the flagellum I may put down 17-19 p-. In the profile view it appears as a fine line, very faintly stained by acid-fuchsin (figs. 40-42). In the surface view of the chamber-wall (fig. 37), it appears in optical section as a glittering dot, which, by varying the focus of the micros- cope, may be made to move continuously away from, or toward, the central refractive spot of each nucleus, as the case may be. As to the collar, my opinion was very uncertain for a long E. MARSHALLI. — FLAGELLATED CHAMBER. 139 time after the flagella had become known to me, although sub- sequent experience has proved that I had really that structure under my observation. F. E. Schulze's paper ' Zur Histologie der Hexactinelliden ' ('99a) gave me fresh encouragement to renew my investigations into the matter, and for this purpose I went once again to Dôketsba in the spring of last year, in order to obtain a new supply of E. marshalli preserved in a number of ways. As before, corrosive sublimate as the fixing reagent gave the best results ; and a careful search on sections stained "with acid-fuchsin, using a very high power (Zeiss' homogene Immersion), resulted in convincing me of the indubitable presence of a collar to each cell. Having once become acquainted with its appearance, I found that it w^as visible in nearly equal clear- ness in many of my old preparations colored with borax-carmine. In order to see them well, the collars must be seen in the profile. The section should be neither too thick nor too thin ; in the latter case it is difficult to recognize the chamber-w^all itself. Moreover, the section of the wall must be so favorably situated that the collars and the flagella stand out in a perfectly clear light, which should not be tinted by the colored light diffused from neighboring parts lying out of the focus. I have never succeeded in perceiving the collars in optical section on the surface-view of the chamber-wall, the difl'use colored light coming from the reticular beams and nodes being sufficient to conceal them. The collar (figs. 40-42) in the profile view appears as a narrow sheath around the base of each flagellum. It is exceedingly delicate, quite clear and very faintly colored by acid-fuchsin. The lateral contour-line is fine or moderately sharp ; the distal edge-line, always very fine. The shape is variable, apparently 140 I. IJIMA : HEXACTINELLIDA. I. owing to shrinkage caused by the action of the reagents. I think it is approximately cylindrical in the natural state. In the preparations that shape is sometimes retained ; but more frequently the collar either gradually narrows toward the distal end or is somewhat narrowed in the middle section, in which latter case the distal end is often expanded in a funnel-like manner. The flagellum traverses the collar either at its middle throughout or along one of its lateral edges after having inclined to that side at a certain distance away from the origin of the flagellum in the center of the distal nuclear surface. In height the collar measures 0-6/7. (5.6 // on an average). The breadth usually measures only IY2-2 // (1-7 /^ on an average), i.e., about as much as the diameter of the nucleus. It may however occasionally reach 3 /^ at the base or at the expanded distal end of the collar. In one or two instances I have seen a line apparently stretching itself between and connecting the flaring rims of a few consecutive collars, which line reminded me at once of Sollas' membrane. But I have satisfied myself that it is to be regarded as something accidentally produced, — possibly a flagellum or portions of flagella laid down upon the free ends of the collars. The collars stand out freely and solitarily, being separated from one another by a comparatively wide space whose width may be said to be on the whole about equal to the distance between the nuclei of the respective choauocytes. Observations of the chamber-wall in the fresh state, 2-5 hours after the capture of the specimens, did not reveal anything of much importance. The preparation of a piece of the fresh choanosome for examination under the microscope necessarily involves more or less dislocation of spicules from their proper E. MAKSHALLI. — FLAGELLATED CHAMBER. 141 positions, wliile the trabeculse suspending the chamber-wall suffer at places unnatural slackening and at other places tightening, or are even wholly broken off. The consequence is that the chambers mostly lose their original inflated form and may even become shriveled up, which is probably to be explained by the inherent contractility of the choanocytal protoplasm. The chamber-wall then does not show the reticular structure ; I suppose the meshes have been obliterated as a result of the contraction. At the best it presents itself as a continuous layer of densely but irregularly granular protoplasm. In optical section it appears somewhat thicker than when seen in hardened preparations. Under favor- able circumstances, the flagellation can be distinctly observed, though no longer in motion. The flagella seem to be somewhat more densely situated than in hardened preparations, which apparently stands in relation with the contracted state of the chamber-wall. I have also seen a number of flagella apparently emanating from little masses of granules in teased preparations. The collar in the fresh state I have not succeeded in detect- ing. This failure was undoubtedly due to the fact that at the time of my investigations on fresh specimens, years ago at the Misaki Marine Laboratory, I had no knowledge of the Hexacti- nellidan collar and moreover no higher power at my disposal than Zeiss' objective DD, which, as I afterwards learned, is by far too weak for the clear observation of the structure in question. It may here be mentioned that in a sketch which I made, in 1895, of the chamber-wall of Acanthascus cactus as examined in optical section in the fresh state, I find the flagellation represent- ed by a series of lines of unequal lengths, the shorter of which are confined to the base in junction wâth the layer of granular protoplasm. It occurs to me as quite possible that I have seen, 142 I. IJIMA : HEXACTINELLIDA. I. in addition to the flagella proper, the lateral contour-lines of the collars and that the shorter lines in the sketch stand for these. I regret that I have had no opportunity to renew my observations on fresh specimens. I have of course not neglected to try silver-nitrate methods on fresh specimens with the view of demonstrating cell-outlines in the chamber-wall. The methods referred to browned the protoplasm but always failed to bring out the expected boundary lines. Not only this negative result, but also the perfectly con- tinuous appearance of the substance constituting the reticular beams, strongly inclines me to believe that the membrana reticularis represents, so to say, a fenestrated syncytial layer, — in other words, that the individual choanocytes stand in organic fusion with their fellows by the aforesaid beams or the lateral protoplasmic processes. And I think this is not a phenomenon that stands quite alone in the group of the Spongida taken as a whole. For, it will be conceded by all that the reticular beams of the Hexactinellidan chamber-wall correspond in all probability to those protoplasmic processes which are known to extend radially from the bases of choanocytes in a number of other sponges. For several species of the Calcarea {Ascetta primordialis, Sycandra rwphanus^ Vosmaeria corticata), E,. v. Lendenfeld ('92) has stated that these processes anastomose and form a network. Sollas ('88, p. XXXVlll) has also made the statement, I suppose for the Spongida generally, that the same continuously unite each choanocyte with its sur- rounding fellows. The structure of the chamber-wall and of the single choano- cytes as described by me in the above unfortunately does not accord in some important points with the description given by E. MAKSHALLT. FLAGELLATED CHAMBER. 143 F. E. Schulze ('99 a, 19'«) of the same in Schaudinnia arctica. Firstly, as to the open or closed nature of the generality of the meshes in the reticular membrane. According to Schulze they should be closed in Schaudinnia arctica, as he originally believed them to be likewise in Eupl. aspergillimi and in several other species ('80, '87), though not in the same way. His original conception seems to have been that the choanocytes lay apart from one another but were joined together by band-like connecting bridges and were disposed in a layer over and upon a continuous basal membrane, much in the same manner as is observed in other classes of the Spongida. His histological study of Schaudinnia arctica, however, led him to the belief tliat such a basal membrane, or in fact any membrane which might delimit the chamber-wall from the incurrent lacun?e did not exist (see p. 137) ; and he thus may be said to have come very near to recognizing what I consider to be the fact, viz., the open state of all the meshes of the reticular membrane. Schulze however entertained quite a different view of the matter ('99a, p. 201 ; 19'«, p. 98). The choanocyte is described by him as having a thin basal expansion — a ' fussplattenartige Ausbreitung des Basaltheiles der Zelle' — which should join with the same of the neighboring choanocytes and form a continuous membrane, called the * Basalplatte.' This should be traversed by the branching and anastomosing granular bauds which radiate from around the nuclei and bring about the reticular pattern visible when looked at on the surface. I make bold to say that the existence of such a * Basalplatte ' is to me exceedingly doubt- ful, but it would be well to leave this moot point to be decided by the results of further investigations. Secondly, as to the prosopyles. Schulze ('80, '87, '99 a, 144 I. IJIMA : HEXACTINELLIDA. I. 19'«) has always represented these to be smooth-edged roundish pores of different sizes, oj)ening through the membranous chamber-wall in variable numbers and in irregular distribution. I should think this amounts to about the same as to say that only a limited number of the numerous meshes of the reticular membrane are open, which in my view should be the case with all. I beg to remark tliat when insufficiently stained, or Avhen subjected to macerating influences, the choanocytal protoplasm appears quite ill-defined as to its limiting contour ; and then, especially if there should be found in the quadrate meshes some granules or strings of granules, — which in reality belong either to the finer branch-beams or to the terminal branches of certain trabecule, and which in other cases seem to be due to dis- integration of the protoplasm, — one may easily be led to think that the meshes are overspread with a transparent film, while here and there may occur such as happen to be exceptionally clean within, but, which being surrounded by a granular tract, may be taken for the only ones that are open. Thv'dly, as regards the general shape of the choanocyte. According to Schulze, it should be nearly cylindrical in the living state and somewhat wine-glass-like when preserved. Dis- tally to the thin basal expansion already referred to, there should follow an elongated neck-like section of the cell-body, which section reminds us of the collum or rostrum of the choanocytes in other sponges. The parts in question in consecutive choano- cytes were observed to be separated by a system of narrow interspaces. Distally they broadened, becoming somewhat trumpet- like, and were finally capped each with a broad collar. A deli- cate central axial-thread extended from the center of the distal E. MAESIIALLT. — FLAOELLATED CHAMBER, 145 surface of the basall y situated nucleus to the origin of the flagel- lutn at the terminal surface of the collum-like section. Now the choanocyte of E. marshalli, as I have described it, lacks the collum-like middle section. It has been described as consisting only of a collar and of a flattened body, which at any rate partially corresponds with the base of the choanocytes in Schulze's sense (1 si\y 'partially, merely because I do not assume the presence of a web-like closing membrane between the lateral processes or reticular beams). In attempting to reconcile the above difference in our stand- points, the following possibilities suggest themselves : 1) The collum-liJce section may he something that really is entirely wanting in E. marshalli. In view of the general uniformity of histolo- gical structure throughout the entire group, the assumption of such a marked variation seems to be scarcely warranted. And yet, a considerable range of variation, so far as the size of the choanocytes is concerned, is to be admitted ; for, Schulze has given for the total height of choanocytes in S. arctica 10-12//, and for the greatest breadth o.Q « ; against which dimensions, the size of the same in E. marshalli as found by me is only about one- half as large or even smaller. 2) The part lohich I have taken solely for the collar may include the collum-lihe section of the cell-body. The narrow shape of that part seems to lend color to this possibility. However, I have never detected the slightest difference in the appearance of the basal and distal parts of the structure in question. It is uniformly and homogeneously trans- parent throughout. 3) / may h.ave seen only the collum-lihe sec- tion, but not the true collar. I must declare myself against this, as well as against the preceding, assumption, on the ground that the flagellum is frequently seen to bend in one way or the other 146 T. TJIMA : HEXACTINELLIDA. I. soon after its origin from the nnclenr surface, and to pursue its course edgewise, either for a part or the whole of the remainder of its passage through the structure in question. This fact indicates that the latter structure must he hollow from its hase. 4) TJie collum-like section may he only a jjortion of the collar. Significant in this respect appear the flattened cake-like form of the hasallj situated nucleus, notwithstanding the elongate shape assigned to the cell-body. It is also to be noted that the collum- like section should have, according to the describer, clear contents, the granules being apparently confined to the region immediately around the nucleus and to the reticular bands in the membranous basal expansion, — a fact which strikes me as somewhat remarkable. And, I venture to say that some of Schülze's figures — notably 19'«, Taf. Ill, Fig. 4, which'shows no material difference in appear- ance between the collar and the coUum, while the angular lateral edge-line between the two parts may be taken as due simply to the presence of the peculiar connecting membrane at that level — are in no small measure suited to give an impression pointing to the possibility, not to say the probability, of the present assumption. However, it will require more substantial grounds to establish tlie point. After all, I am inclined to give weight only to the first and the last of the four above-mentioned possi- bilities in the way of explaining the position I have taken as regards the structure of the choanocyte in E. marshalli. Fourthly, an interesting discovery was made by Schulze in S. arctica in that the closely standing distal ends of the collum- like section of the choanocytes, at the boundary between them and the collar, were connected laterally with one another by a cementing mass, though at times they appeared there to be simply sticking together. In the plate-like connecting mass sur- E. MARSHALLI. — TRABECULE. 147 rounded by three or four adjacent cells, there were sometimes found roundish pores, which should allow the entrance of water into the chamber. In E. marshalli I have seen no trace what- ever of this kind of connection between the cells. It is probably a sort of an accessory arrangement which is not of universal occurrence among the Hexactinellida. The trabecule. — These are in general fine and thread- like, in places band-like. The disposition of their branching and anastomosing strongly reminds one of the irregular web woven by certain spiders (tr., PL IV, fig. 22 ; PI. V, figs. 36, 43 ; see also PL VIII, figs. 29, 30). The meshes are of quite indefinite shape and size. Generally speaking, the trabecule stand for the mesenchyme of other sponges, but never and nowhere do they form a bulky mass or a compact layer of any considerable thickness'", though in places they may be expanded into film-like membranes. Such expansions occur here and there in the deeper parts, at points where three or more trabeculse join together, and are in appearance somewhat alike the nodal confluence of the filamentous pseudopodia of certain Rhizopods. The same membranous development of the trabeculfe, but on a far greater scale, is seen in certain definite positions, especially on the surfaces delimiting the sponge from the exterior. Her.e belong the dermalf, the gastral and the canalar membranes, as *The diagrammatic figure, recently given by Delage and Hérouaed in the 'Zoologie Concrete ' (T. II, PI. 8, fig. 4), representing the relation of parts in tlie wall of EuplecleUa, is fitted to give an altogether erroneous idea of the structure, in that the choanosome is shown as a tliick folded layer of compact ' mésoderme ' inclosing the chambers, wliile botli the ectosome and the eudosome arc given likewise as thick, minutely perforated layers con- nected with the choanosome by solid pillars. Their other diagrams on Pl. 11, relating to the Hexactinellida, are much better. But these, us also the matters embodied in the text, do not call for special couniient, being entirely based on the representations of F. E. Schulze. t For the fact that this layer in E. marshalli hardly deserves to be called a membrane, on account of the general thread-like development of its beams, see ante, p. 120. 148 I. IJIMA : HEXACTINELLIDA. 1. well as the oscular membrane, the marginal membrane and the membrana reuniens, all of which have been s]:)oken of before as adaptations of the general trabecular system. I will repeat that all these membranes are not different from the thread-like trabecuke either in histological character or in the manner of their occurrence. I am even under the impression that, during life, there obtains in the tissue concerned a certain degree of instability in the form, whether membranous or filamentous, which is assumed at different times. As the result of a certain stimu- lus, causing protoplasmic contraction, a membranous area may thin out and finally break apart in the middle ; then, by en- largement of the gap or gaps thus produced, the area may readily convert itself into a mesh or a series of meshes bounded by filamentous beams (see PI. VIII, fig. 30). Contrariwise, the trabecular cobweb may become so close meshed as to finally fuse together into a continuous sheet, or a part of it may be so drawn out as to form a filmy expansion. Indications of such trans- formations are indeed very frequently to be observed. This theory presupposes a viscous semi-fluid nature of the trabecular substance, which assumption seems also to explain the apparent facility with which floricomes traverse a dense cobweb of trabe- cule in order to reach the tips of the dermal hilt-rays. The above nature appears to me all the more assumable, since, as will soon be dwelt upon at length, I am inclined to ascribe no pinacocytal covering to the entire trabecular system, but to regard this as a network of bare-surfaced syncytial protoplasm. A spicular sheath, consisting of a continuous layer of the soft tissue, has been assumed or mentioned by some writers (Thomsox '70, p. 710 ; Schulze '87, p. 24). Although I have never been able to prove the fact, yet I can not but hold it very E. MAKSHALLI. — TKABECULJE. 149 likely that the spicules, during their growth or the deposition of new siliceous matter over their surface, are covered uniformly all over by an excessively thin layer of the matrix. On the other hnnd, the impression I have repeatedly received from the observation of the larger parenchyma lia in well-colored preparations, has been that these have no other coating than a layer of an irregularly meshed trabecular network, lying in direct contact with the spicular surface. It is not at all improbable that many old spicules, though in situ within the bounds of the sponge-wall, are to be considered as lying partially outside of the soft parts, as the prosials as well as the rhaphides and the floricomes at the tips of the dermal hilt-rays undoubtedly do. In the fresh state (2-5 hours after capture), I have ob- served the trabecular substance to be either simply minutely and densely granular or composed of a clear homogeneous ground-sub- stance inclosing a greater or less quantity of opaque and irregular granules. The nuclei presented themselves as refractive spherules. Nowhere on the surface was flagellation observed. Nor have I been able, notwithstanding my special endeavors with silver- nitrate and methylenblau methods, to bring out cell-outlines either on or in the trabeculse. When carmine particles were added to the fresh preparation, they stuck to the trabecular sur- face with a certain degree of firmness, so that they could not be moved by the water current produced under the cover-glass by the use of a blotting-paper. Watching such preparations attentively or viewing them at brief intervals under the microscope, the while keeping them perfectly quiet, the attached carmine particles, and no less the nuclei and the protoplasmic granules, were seen for some time slowly to change their relative positions, which change was accompanied by a slight alteration in the form 150 I. IJIMA : HEXACTINELLIDA. I. of the trabecula itself. This indication of life, which I have observed in E. inarshalli as ^vell as in Acanthascus cactus, I interpret as due to protoplasmic contractility. As for a flow of protoplasm, it is not possible, since the trabeculse, though soft, pliant and easily destroyed by pressure, are much too consistent to be with any exactness called fluid. Observed iu hardened preparations, tlie substance of the flUimentous trabeculœ is tolerably well stained and dense-looking, being either nearly homogeneous or granulated in varying degrees. The membranous expansions, looked at face on, usually present a somewhat less dense or clearer appearance, apparently due to the thinning out of their substance in forming such areas. Here are seen granules and irregular particles or little streaks, which are scattered either somewhat uniformly or in a manner suggestive of a reticular arrangement (PI. IV, fig. 23 ; PI. VIII, fig. 30). This appearance may be partially the result of the hardening process. A fibrous or streaked appearance is also not infrequently noticeable, generally running parallel with the free edges of the band-like or otherwise membranously developed trabecule. In most such cases, I have convinced myself of the fact that the appearance is due, not to the real existence of differentiated fibers, but to fine wrinkles which are probably produced by contraction ; for, the streaks are not only of very indefinite contour in certain places, but also at the ends are seen gradually to lose themselves in the membrane as it generally ap- pears. The edge-line of the trabeculse is always of simple contour. Sometimes unusually coarse, réfringent and well-stained granules are found in isolated occurrence. These are j)robably the same as those found inclosed in the thesocytes to be described further on. E. MARSHALL!. — TRABECULyE. 151 The trabecular nuclei (PL IV, fig. 23; PI. V, figs. 36, 43, le- staining processes has never resulted in demonstrating the pre- sence in it of parts with any ditïerence in the power of selecting 152 T. TJIMA : HEXACTINELLIDA. I. stains. It will then require no further words to explnin that I regard the tnibecuhie as consisting of the fused cytoplasm of the cells represented by the above free nuclei. Here I leave this point, to resume it soon again. The question whether the trabeeuhe have a pinacocyte covering or not, I am fully aware, is a delicate one. F. E. Schulze has always assumed its presence in the forms studied by him, an assumption which seems quite justifiable from a theoretical point of view. However, as before indicated, I have been led to the contrary opinion, though at first I felt much difiidence in coming to this conclusion. Mention has already been made (p. 33) of my failure to demonstrate cell-outlines on the trabecular surface. The history of our knowledge of the pavement epithelium in the Spongida teaches us caution in drawing conclusions from that negative result ; but in the present case, I am quite at loss to believe that my methods were in any way at fault. Besides, there is another circumstance, which, simple as it is, seems to me to deserve due consideration. It is the excessive thinness in which the trabecule, whether filamentous or membranous, so often present themselves in all parts of the sponge- body (see PI. V, fig. 43, ii\', also PI. VIII, fig. 30). The thinness is such that barely enough room is given for the protoplasmic granules to arrange themselves in a single row or layer, as the case may be, and this seems to be scarcely compatible with the assumption of the plurality of differentiated tissues or layers. To all appear- ance, the thinner trabeculre are nothing more than simple threads or films of the protoplasm. In the absence of indications to the contrary, I see no ground for hesitating to ascribe the same E. MARSIIALLT. — TRABECULiE. 153 structural conception to the relatively thicker parts of the trabeculœ. To be plain, my notion is that for once among the sponges we find in the Hexactinellids that the development of pinacocytes, both as an investing of the exterior and a lining of all the internal cavities and passages, is entirely suppressed. This seems to be in harmony with a certain point in the structure of the chamber-wall as described by me, viz., that the convex outer (incurrent) surface of this is, so to speak, naked, the basal ends of the choanocytes being directly exposed to the water, — a fact which I tliink is nearly, if not quite, admitted also by F. E. Schulze Çgga, p. 209 ; 19'«, p. 98), in that the existence of a basement membrane at the phice is denied by him and that of a pavement epithelium held doubtful, while the choanocyte layer is considered as resting on a relatively ivlde-meshed net- work of trabeculœ. Be that as it may, F. E. Schulze's rej^resentation of certain nuclei and cells as pinacocytes seems to be open to discussion. In E. aspergillum ('80, pp. 669-671 ; '87, pp. 23, 24), he distinguished the three following kinds of nuclei or cells in the trabeculœ : 1) Small, spherical nuclei, abundantly and tolerably uniformly scattered. These were seen on profile view to project a little above the general surface of the trabecule, and were thus con- sidered to occupy the most superficial position and on that ac- count to represent flat epithelial cells, whose outlines were certainly not seen. The nuclei in question evidently corresponds — in part at least — to those which I have called the trabecular nuclei. It is then important to decide if their seat in the trabeculœ is really superficial in relation to that of certain other nuclei or cells in the same. This is by no means so, to judge 154 T. IJIMA : HEXACTINELLIDA. I. of what I have seen of the same nuclei in E. niarshalli and in a number of other Hexactinellid species, the histology of which will be duly remarked upon in the future numbers of this series of Contributions. On the contrary, the occurrence on the syn- cytial trabecular thread, of certain cells, which on account of their well-defined spherical or ovoid bodies cannot possibly be pinacocy tes, is quite common. The projected state of the trabecular nuclei, as they have come under my observation, was in nearly all cases apparently due to the simple fact that the space in the protoplasm was not sufficient to completely include them. The unilateral situation of the nuclei, especially in the thinner trabecular thread, may also be considered as the result of an uninterrupted and most effective bringing about of protoplasmic continuity, vvhich may be of primary importance to the trabecule considered as a connective substance. 2) Somewhat larger and more oval-shaped nuclei with little protoplasm around them, giving simple stellate or spindle-like shape to the cells. These appear to have been seen in relatively smaller numbers and are regarded as representing connective- tissue cells lying in a hyaline matrix. It occurs to me likely that the 'nuclei' here referred to, correspond in part at least to certain distinct cells, which I will describe in a later chapter under the designation of arch^eocytes. My grounds for holding that opinion will be given presently in connection with ap- parently the same ' nuclei ' described by Schulze from Schau- clinnia arciica. Tlie nature of the protoplasmic space around the alleged nuclei, seen by Schulze in E. aspergillum but apparent- ly not in S. arctica, remains incomprehensible to me, unless it be considered to be a part of the protoplasm of the trabecular E. MARSHALLI. — TEABECüL^. 155 syncytium having especially concentrated granulation or being otherwise exceptionally modified. 3) Larger cells, likewise considered to be situated in the hyaline connective-tissue matrix and distinguished by a more or less abundant nccumulation of refractive, inten.sely stained granules of various sizes. These granules were occasionally of a brownish or yellowish color. They have been compared, quite justly I think, to fat or starch in the physiological sense. I believe that this kind of cells are of very general occurrence in the Hexactinellida, and I take it to be analogous, and in all probability homologous too, to the thesocytes (Sollas) of non- Hexactinellids. (See anon under Thesocytes). Let us now compare the above with the results arrived at by the same investigator in the case of Schaudinnia arctica ('99 «J PP- 206-209 ; 19'a, p. 98) and connect therewith such remarks as may seem conducive to a clear understanding of the matter. In that species, the nuclei or cells distinguished by Schulze in the trabeculœ are essentially the following two : Firstly : cells containing a mass of j^eculiar spherules and conglomerates (' Knollen ') around the nucleus, and which are numerously present on the dermal and the gastral membrane, on the thicker subdermal and subgastral trabeculse, and around as well as between the apopylar openings of the chambers. On account of the bulky contents, they project more or less over the surface of the said parts in a hump-like manner. These cells are taken by F. E. Schulze for pinacocytes. However, from the nature of the 'Knollen' described ('99a, p. 207), it is exceed- ingly probable, and indeed scarcely to be doubted, that we have here to do with the thesocytes above referred to under (3), — cells, which, unless I am greately mistaken in my homologization. 156 I. IJIMA : HEXACTINELLIDA. I. Schulze had before in E. aspergillum associated with the con- nective substance, instead of classing them as flat epithelial cells. I am decidedly in favor of this older view of Schulze's. The cells in question, as known to me from Euplectella and a number of Hyalonematids, Rossellids, <&c., — which cells I have no doubt belong to one and the same class of cells both morphologically and physiologically, — always have plump bodies and show clearly defined cell-boundaries, lohich as a rule are well separated from one another. In most species, e.g., E. marshalU, they are ordinarily quite sparingly and isolatedly present, while in certain places and under special circumstances they may occur in large, compact masses (PI. IV, fig. 24). (See anon under Thesocytes). These facts in my view militate against the propriety of attri- buting to the cells an epithelial nature, in spite of their most superfi.cial situation. Even granting their epithelium-like arrange- ment, their homology with the true pinacocytes may be questioned, because the latter cells, so far as my knowledge of them goes, are never known to have similar contents, Avhile certain other cells which do possess such contents — i. e,, the thesocytes, which are not improbably identical cells in all sponges of different classes — are found scattered invariably in the mesenchyme of such sponges as show besides a true pinacocytal epithelium. It is important to note that Schulze did not find in Schaudinnia arctica the alleged epithelial cells on all parts of the trabecular systems. They were evidently not present on the finer intercatneral, subdernial and subgastral trabeculœ. " Es könnte daher sein, dass diese Verbindungsbalken einer besonderen epithelialen Bekleidung entbehren und ganz aus der...Biudesub- stanz bestehen " ('99«, p. 208) is his conclusion. While main- taining the presence of an epithelium on the thicker trabeculœ, E. MARSHALLI. — TRABECULE. 157 the possibility and perhaps the probability of its absence on the thinner ones is thus admitted by him. Such a partiality, if true, must be said at least to indicate a very remarkable case for an epithelium ; whereas, if we had to do wàth cells of mesenchymal nature, there would be nothing extraordinary in finding these attached only to such thicker beams as are capable of bearing them on. Secondly ; nuclei without the ' Knollen ' around them, and which are numerously found, apparently situated in the connective- tissue substance of the trabecular threads and membranes. These nuclei are said to be on the average somewhat larger and of a more oval form than those of the cells with the ' Knollen,' though in some parts of the trabecular system there are present a number of such as are indistinguishable from the latter nuclei in point of size or of other features. In view of this agreement sliowai by the relatively smaller nuclei under consideration, Schulze ('99a, p. 208) leaves open the possibility, if I understand him aright, that these smaller nuclei may belong to flat epithelial cells w^hich have not accumulated the * Knollen ' in their bodies. The relatively larger ones are at any rate regarded as the connective-tissue nuclei. No special protoplasmic space around them seems to have been seen, for no mention of it is made. These nuclei are said to be the only kind that are found in the thinner trabeculœ and especially in those spanning the spaces between the chambers ; so that, to follow the writer's expression as nearly as possible, * here at least ' the existence of an epithelial covering could not be ascer- tained {I.e. J p. 209). However, he expressly leaves undecided (19'a, p. 98) whether or not the diÔerence in size of the nuclei ' alone and in all cases suflices for the distinction of the two kinds of cells ' (i.e., flat epithelial cells and connective-tissue cells). 158 I. IJIMA : HEXACTINELLIDA. I. On the whole, these ideas of Schulze concerning the nuclei do not differ from those expressed by him before with regard to the smaller and the larger nuclei in the trabecuhe of E. asper- g ilium, mentioned and remarked upon respectively by me under (1) and (2) on pp. 153-154. This might naturally follow, if, as I have assumed, the cells with granular contents in E. aspergillum, mentioned by Schulze and referred to by me under (3) on p. 155, are the same kind of cells as the ' Knollen ' cells of Schulze (thesocytes) ; for, in that case, the nuclei without the ' Knollen ' in 8. arctica would have in E. aspergillum nothing to correspond to but the nuclei I have mentioned under (1) and (2). A new point is that the cells represented by the ' smaller nuclei ' in the trabecular, or by at least some of these nuclei, are possibl}^ iden- tical with the * Knollen ' cells or (according to my interpretation) thesocytes. The implication is that one of these two kinds of cells may be directly derived from the other. Tliis seems to me highly improbable ; for, I think I have grounds to believe that the thesocytes arise, on the contrary, from the so-called ' larger nuclei ' of Schulze. In attempting to establish my position here taken, I will begin by stating that I can not help entertaining a doubt as to whether the size given by Schulze for the ' larger nuclei ' found in the trabeculœ refers to real nuclei, — if a certain mistake is not here involved in spite of his wonted accuracy in observations. According to him, they should be 3-4 ij. large (against ca. 2 // of the nuclei of ' Knollen ' cells ; 19'«, p. 08). To S2:>eak from my own experience, the nuclei in the trabecular irrespective of their having a well delimited cell-body or not, are all tolerably uniform in size and general appearance within the limit of the same Hexactinellid species at least, if not of the entire class. They E. MAESHALLT. — TKABECUL.E. 159 are always very small, fiuctuating within only an inconsiderable range of variation in this respect ; in shape they are constantly spherical or approximately so. In E. marshaUl they rarely, if ever, exceed 2 /'- in diameter, while the smallest measure about l\'2/^5 — a variation of say less than half a //. This does not quite agree with Schulze's statements. However, this fact alone would not perhaps have led me to the above skepticism, were it not for another circumstance which serves to account for the size ascribed by Schulze to the ' larger nuclei.' As I have before indicated (p, lö4, under 2.), I hold that the ' larger oval nuclei ' Schulze's in the connective substance of both E. aspergUlum and S. arctica are nothing else than, or should at least include among them, the well-defined cells which I call the archœocytes (PL Y, figs. 36, 39, 43 ; arch.). For the nature and characteristics of these cells, the reader is referred to the special chapter devoted to them. Here suffice it to mention that these cells are exceedingly liable to be taken for mere nuclei, and that their distribution in the trabecular system agrees on the whole with that ascribed by Schulze to the ' larger nuclei ' in S. arctica. On the other hand, I find no mention of these cells as such in the descriptions given by that investigator, though they have been undoubtedly seen by him. At all events there can be no doubt whatever that the ' larger nuclei ' seen by him in groups in the immediate proximity of the flagellated chambers and taken for possible genital cells ('ig«, p. 99),'-' are the same as my archœo- cytes, which thus seem to have been only partially recognized in that they were errorneously considered as nuclei and not as com- plete cells. If this be so, then the discrepancy in our observations * The same groups of cells, likewise considered as possibly concerned with reproduction, were also seen before by Schcjlze in Farrea occa ('87, p. 285). Here the elements in the groups are described as ' cells,' with nuclei which stain with special readiness. 160 T. IJTMA : HEXACTINELLTDA. I. concerning the size of the nuclei in the trabecala? Avill explain itself. It is also plain that, if there prevails a general uniformity of size and appearance among all the nuclei, which I hold to he the case, the assumption of the identity of this or that kind of cells with another should lose all its validity, so long as it is founded on the similarity of their nuclei alone. Finally I will mention here, reserving the details of my observations relative to the point to another chapter, that the thesocytes seem to he one of those kinds of cells which arise directly from the archreocytes, the transformation being effected simply by gradually accumulating the 'Knollen' in the cell-body (PI. IV, fig. 24). It may perhaps be urged that the archseocytes may be present in some species with, and in others without, the cell- outlines, and that the size of the nuclei may likewise be a vari- able matter ; but the fact would remain the same that the theso- cytes— the 'Knollen ' cells of S. arch'ca included— develop out of the ' larger nuclei ' of Schulze. To resume my ir, ofthe Osaka Mercantile Steamship Company, has informed me that in liis native province, Suwö, Euplectellœ are well-known objects, being considered as indis- pensable to the marriage ceremony (see anon under Misc. Notes), and that lie used to think that they were fished up near Öshima, close to the coast of that province, in the Inland Sea and not far distant from Siiinionoseki. E. OWENI. — GEN. CHARACTERS. 205 Spec. Total length, excl. of basal tuft. Breadth at tlie broad- est part, ledges iucl. REMARKS. A mm. 110 mm. 16 Mr. Ko]MEYAM.\'s .specimen. B 117 24 X 21 JJ J? ,v C 120 35x30 Measurement given by Schulze ('87). D 138 35x28 Mr. Kojieyama's specimen. E 152 35x31 Sei. Coll. Mus.; près, by Mr. Öno. F 160 t30 „ „ „ „ ,, „ MiMATS. G -165 t41 ?; 7j jj jj j> j> uno. H 180 34x30 PI VI fi-^«- 1. I 180 45x35 Mr. Komeyama's specimen. J 204 34 Mr. Owstün's specimen. PI. VI, fig. 4. K •••••220 t37 Sei. Coll. Mu?.; près, by Mr. Nakagawa. L 220 40 Mr. Komeyama's specimen. M 226 37 Measurement given by Marshall ('75). N 240 t45 Sei. Coll. Mus.; près, by Mr. Owston. 0 -250 t40 „ „ „ „ „ „ Nakagawa. P 311 46 Measurement given by Marshall ('75). Q 360 62 X 48 Sei. Coll. Mus.; près, by Mr. Öno. PI. VI, figs. 2 &3. * Sieve-plate wanting, and therefore excluded in measuring the length, t Breadth measured after restoring the collapsed body-wall to a cylin- drical form. The ratio of body-length to the greatest breadth may be given at 1: .15 — .25. As compared with 1: .3 — .44 of E. mar- shalli, the present species must be said to have in general a distinctly more elongate shape. Both Marshall and Schulze have stated that the body is more or less compressed laterally, presenting an oval form in cross-section. This is no doubt generally true. However, we 206 I. IJIMA : HEXACTINELLIDA. I. have here evidently a feature much subject to iudividual varia- tion. Perhaps, as a general occurrence, the compression becomes gradually more and more marked as the sponge grows in size. Sometimes specimens of considerable dimensions may be approx- imately circular in circumference for nearly the entire length ; such a case is found in Spec. L (of the above list), which, being preserved in alcohol and in good condition, has undoubtedly retained the natural form. As a rule, the upper terminal region of the body may be said to have a roundish form in cross-section. Spec. J (PI. VI, fig. 4) is exceptional iu so far as the greater part of the body is nearly cylindrical, being almost circular in cross-seclion, while the upper end is perceptibly flattened, here the breadth measur- ing 30 mm. in one direction and 25 mm. in the other. Constant seems to be the pronounced compression of the body-wall at the contracted inferior extremity, where the basal fibers are given off. To give the cases in which I have measured the diameters at this position : Spec. A 8x 5 mm. B lOx 7 „ I) 12x 6 „ I 17x11 „ /. 18x10 „ Q 26x12 „ The compression at this end seems to be independent of that in the upper main portion of the body, for I have found the i^lanes of the two not always exactly coinciding. They may be disposed even nearly vertically to each other ; so, e. g., in Spec. /. E. OWENI. — GEN. CHARACTEKS. 207 The compressed, inferior end is normally closed by a bottom- plate. The apparent absence of this in some specimens is probably due simply to damage after capture. It is of essentially the same appearance as that of E. marshalli. The same holds true of the strongly arched, often hemispherical sieve-plate of the superior end as well as the portion of the basal tuft nearest its point of origin. The occurrence of the bottom-plate and the condition of the basal tuft indicate that the mode of insertion of the species into the substratum is likewise the same as in E. marshalli (p. 93). In many specimens the basal tuft is present as a clean, silky lock of considerable length. As more normal are to be considered the cases in which I have found the tuft form, a short distance below the point of its origin, a bulky, irregular or elongate mass, including a copious quantity of sand, shell- fragments, worms-tubes, &c. Parietal ledges and the cuff have been hitheito considered to be entirely wanting in the present species. I must say that such is not always the case ; in fact, both the structures men- tioned seem to be of common occurrence, though they are never so prominently or so extensively developed as in E. marshalli. Here again it seems we have to do with a character which is subject to considerable variation according to individuals. The beautifully preserved specimen I have shown in PI. VI, fig. 4 (/ of the list on p. 205) approaches most closely to the descriptions given by Marshall and Schulze in respect of the character of the external surface. In it, the parietal ledges are at most simply suggested, the interspaces between the parietal oscula having in general a gently convex, external surface. 208 I. IJIMA : HEXACTINELLIDA. I. Whereas, all other specimens before rae (exceptuig one doubtful case with much abraded surfaces) show a greater or less number of ridge-like prominences or ledges, such as are fairly well exemplified in PI. VI, figs. 1-3. Even the smallest specimen (.4) exhibits a decidedly uneven surface, somewhat as in Pi. IV, fig. 9. The ledges in the larger specimens may be 5 mm. high but are more usually much lower. Their free edge may be said to be tolerably even ; it is either blunt or sharp. In length the ledges are quite indefinite, often rather short. They run in the usual, irregular manner, but generally in either transverse or oblique directions. In a certain specimen (Spec. /) I have found the ledges for the most part somewhat unusually sharp-edged and supplied along the edge with an inconspicuous, palisade-like row of spicules, projecting to a length of about half a millimeter. In the case of the more blunt-edged ledges of the same specimen this palisade was wanting. Nor have I noticed it on any ledges in all the rest of my specimens. On close observation it was seen to consist of spicules similar to those which were likewise inconstantly found on the sharper-edged lappets of E. mar- shalli (p. 97). The cuff is on the whole inconspicuous, especially so in the smaller specimens. It is quite usual that different parts of the sieve-plate circumference show the cuff in different states of development. It may in places even be wholly wanting. In its highest development, the breadth does not exceed 4 mm., as measured on the upper surface. It is of moderate thickness and sharp-edged, being directed either outwards or more or less up- wards. Generally there exists no spicular fringe along the edge ; E. OWENI. — GEN. CHARACTEKS. 209 sometimes, however, spicules of quite inconspicuous length may project out here and there. The line of insertion of the cuff, i. e., the juncture-line of the lateral parietes with the sieve- plate, as seen from the side, is generally irregularl}^ wavy. The parietal oscula measure 2 mm. or less in diameter. They are arranged on the wdiole with tolerable regularity in longitudinal and transverse rows (PL VI, figs. 1, 4). Here and there, this regularity is subject to disturbances, conditioned in a measure by the development of, and the course taken by, the parietal ledges. Thus, by the sides of an obliquely running ledge it is usual to fiud the oscula arranged in rows running parallel to it. In some specimens (e. g., Spec. Q, shown in figs. 2 & 3), the distribution of the oscula may be said to be generally rather irregular, which fact may stand in relation to the wide-spread occurrence of ridge-like elevations over nearly all of the external surface. With regard to the appearance of the parietes on the inter- nal side and to the arrangement of beams in the skeletal frame- work, what I have recounted for E. marshalll may be said to be essentially applicable to the present species also. Marshall ('75) had described the occurrence of both the circular and the longitudinal skeletal beams in sets of twos run- ning side by side — such as might arise by the splitting lengthwise of every, originally single beam — as somewhat constant and characteristic of the species, which generalization has however not been fully borne out by facts subsequently brought to light. F. E. Schulze ('87, p. 79 ; '95, p. 30) has found in the small specimen examined by him (Spec. C of the list on p. 205) that 210 I. uni A : hexactinellida. i. tlie peculiarity referred to in the arrangement of the beams oc- curred only here and there without regularity in the longitudinal system, and as regards the circular beams, only in the upper region of the body. My observations are in general accord with Schulze's, The arrangement of the longitudinal beams is exactly similar to that observed by Schulze ('95, p. 25) in E. simplex and by me in E. marshalli (p. 94). The same may also be said with respect to the circular system. Only I have to add that the relatively close arrangement of the circular beams noticed by Schulze in the upper region of the body is to be observed only in the younger specimens in which that region is still actively growing, and the said beams are there either under- going, or have comparatively receîitly undergone, multiplication by splitting. In a specimen of 138 mm. length (Spec. D), I iiave found the region near the upper end still characterized in the way indicated. In all the larger speciuiens, the circular beams are set well apart from one another, notwithstanding the occasional occurrence of anastomosis. After the specimen has nearly attained full size, a number of the uppermost circular beams seem to deviate from their regularly transverse course and become more oblique and wavy, so that they often anastomose and even intersect one another. At their juncture with the sieve-plate, they are frequently seen to be prolonged, like the longitudinal beams, into the beams of that plate, similarly as described by me for E. marxhalli (p. 94). The number of skeletal beams has been counted in five specimens, as follows : E. OWENI. — SPICULATION. 211 Spec. Niunber of circular beams. Nuiiibei- of lonsjit. Ijeams at tlie upper end. Number of lougit. beams at tlie middle. Number of longit. beams at the lower end. A 49 36 25 18 D 54 35 32 23 E 45 p 31 25 L 50—56 35 35 29 Q 65 p 44 28 I claim DO more tliaii approximate correctness for the figures ill tlie above table. In specimens E and Q, instead of the number of longitudinal beams at the upper end, I have counted that of the sieve-plate beams arising therefrom ; these were found to number 38 and 49 respectively. The majority of the meshes of the skeletal framework are perforate, that is to say, they each inclose a parietal osculum. The so-called interstitial or imperforate meshes occur, several in succession one behind the other, between any two, relatively closely situated, longitudinal beams ; they also occur isolatedly, without any regularity as to their distribution. The spicules are always and everywhere free. I find this to be the case even in the largest specimen (Ç) before me. Spiculation. Under this head, my studies go essentially to confirm our previous knowledge as derived especially from F. E. Schulze'« works ('87, '95). Moreover, there exists no marked difference between the spiculatiou of this species and that of E. marshalli or E. iviperialis. I may therefore be brief in my account. 212 I. IJIMA : HEXACTINELLIDA. I. The large oxystauractin-principalia of the circular and the longitudinal skeletal beams may measure 45 mm. in length of the longitudinal axis and 90 /^ in breadth of rays near the center. Other spicules of the beams are almost exclusively thetactins of the usual shape, rarely diactins and paratetractins. The thetactins and diactins are occasionally sufficiently large and strong to be classed with the principalia ; all the rest are thin, ranging from 7 to 20 /^ in thickness near the center. The parenchymalia of the loose tissues are again predomi- nantly thetactins, quite variable in size, running partly in strands and partly in more or less diffuse arrangement. Occasionally there occur paratetractins, rarely stauractins and pentactins, especially among the larger parenchymalia. Many of the comi- talia in the strands are thin and rather short diactins, provided with four tubercles at the center. The oscularia (PI. VI, fig. 10) are predominantly diactins, which have been very aptly called by Marshall compass-needle- like. Length 200-600 /^ and over ; breadth near the middle 6-17 ix. The center usually with two or four oppositely placed tubercles. Tiie oscularia are densely crowded, with their long axis disposed paratangentially to the edge of the oscular mem- brane. In the ring-like zone occupied by them, the innermost are generally the smallest. The outermost are the largest, and some of these may have one or more of the central tubercles produced into shorter or longer lateral rays, thus assuming the form of thetactins, tetractins or pentactins and even hexactins. These lead over the oscularia on the one hand into the paren- chymalia, and on the other, into the gastralia. E. OWENI. — SPICULATION. 213 The basalia (PL VI, fig. 9) differ in no way from the same of E. marshalli, except in being slightly more slender and in having perceptibly smaller anchor-heads. The anchor-teeth, of which there are 3-7 in each head, are strong and abont 40 /i long. The distance from tip to tip of any two oppositely situated anchor-teeth measures 70-85 //.. The entire head is of about the same leuo;tli. The shaft is less than 20 n thick close to its ori2;in from the head, only about 7 /y- at a short distance above the axial cross, and not more than 30 n in the thickest part farther above. — Some abnormally formed anchor-heads that I have found are figured in Pi, VI, figs. 7 & 8. In one of these cases the teeth are developed only on one side of the miter-shaped knob ; the suppression of the development of teeth on the other side is evidently due to the head having lain with that side pressed against a compact bundle of its fellows. — I have discovered no more pentactin-anchors than Schulze did. The dermalia may be nearly 1 mm. long. On the whole they are somewhat smaller than in E. ynarskaUL All the rays are nearly smooth throughout and tapering, but usually bluntly pointed at the free end. Distal hilt-rays mostly 90-130 />« in length and 3-7^/2 « in breadth near the center. Paratangential rays 110-120 /y. long. — Exceptionally and then only along the edge of especially sharp-edged ledges, the dermalia may be of unusually large size. The hilt-ray may here reach a length of 400 /y.. It participates, together with slender diactins, in the formation of the inconspicuous row of bristles, before mentioned as having been found in a certain specimen. The gadralia and canalaria are pentactins showing the rudi- 214 I. IJIMA : HEXACTINELLIDA. I. ment of a sixth, proximal ray in the form of a small protu- berance. Paratangential rays 85-115 n long and 3-7 /^ tliick near the center ; smooth nearly all over. Distal ray somewhat longer, sometimes considerably so ; in the larger gastralia it may be sparingly beset with small prickles near its distal end. Of the hexasters, the floricome measures 75-88 jx in diameter. The oxyhexaster (PI. VI, figs. 5 & (3) is smaller than that of E. 7)iarshaUl, measuring 50-60 /i, sometimes up to 70/>«, in diameter. It is present in abundance, — decidedly much more so than in either E. imperialis or E. marshaUl. Compared with the same in either of these species, both the principals and the terminals are somewhat more slender. The principal is 7 fj. long (as measured from the center of the axial cross) and 2^ jo n broad at the middle. It bears usually 3-4, rarely 2 or 5, divergent terminals at the outer end. The graphiocome was recognized to be present in the species for the first time in '95 by F. E. Schulze, although its termi- nals— the rhaphides — were known long before. I find it is common. It may measure 245 i-^ in diameter, the rhaphides being 114 n long, when fully developed. The latter, after they have fallen off from the principals, are still found as usual in the superficial region of the sponge-wall, though not in such great abundance as in E. iiiarshalli nor in such regular arrangement as has been ascribed to their sheaves by Makshall. The sieve-plate shows the parenchymalia (principalia and comitalia) mainly consisting of thetactins and diactins. The former seem to furnish the principalia more often than do the latter. Occasionally stauractins may occur among the parenchy- E. OWENI. MISCELLANEOUS NOTES. 215 malia. For the rest, the spiculation of the pLite is essentially the same as in other species. Miscellaneous Notes. In PL VI, fig. 2, I call attention to the presence of a small, accessory sieve-plate on the side of the sponge, at some distance from the normal, terminal sieve-plate. A similar case of abnormality has also been noticed under E. marshalU. Another observation, which I should mention in this con- nection, is that once in a specimen {L) of E. oiveni the sieve- plate was found to have an unusually irregular outline, and seemed in certain places to have appropriated the adjoining parts of the lateral parietes by converting the skeletal beams, as these are usually arranged, into sieve-plate beams. The above cases of abnormal development are, I think, of interest, as demonstrating the fundamental unity of the sieve- plate with its angular meshes and the lateral w\all with its round, parietal oscula. The Crustacean inmate of E. oiveni is Spongicola venusta DeHaan, the same as in E. marshalli (see p. 201). The type of that Crustacea, described by DeHaan in the Fauna Japonica, was probably taken from the specimens of E. oiveni, which were taken to Eurojie by Major v. Siebold. Of all the specimens of E. oweiii I have examined, seven possessed each a pair — invari- ably a male and a female — of the Crustacea. The others con- tained some one, some none at all ; but since the sponge-wall was more or less damaged in all of these cases, loss of the in- mate may possibly have taken place in certain instances. 216 I. IJIMA : HEXACTINELLIDA. I. The comparative facility witli which specimens of E. oioeni could be got b}^ purchase or otherwise in Fukuoka, Shimonoseki, &c., is undoubtedly due in a great measure to the fact that they are in some demand among the folks in those parts of the country, on account of an old custom connected with- their mar- riage ceremony. The custom consists in including a Euplectella among the articles with which the room of the ceremony is decorated, or which are taken by the bride to the bride-groom's house. It is held to be a felicitous object betokening eternal connubial love on account of the presence of the inmates in an inseparable pair. In the long list of gifts, which the present Emperor and the Empress of Japan received from their subjects on the occasion of their 25th wedding anniversary, are men- tioned several Euplectellie, gifts humble in themselves but full of well-wishing sentiments. The Japanese name for Euplectella, Kai-rö-dö-kets (written {§" ^ [P] /C), means, as was correctly point- ed out long ago by Marshall ('75), something like ' Together unto old age and unto the same grave.' Perhaps the name may have seemed to the Japanese mind all the more appropriate, since, by simply changing the first of the four ideographs into one which means ' the sea ' (ï^) and yet without changing at all the pronunciation of the entire combination, the name may be made to signify * Lobsters in the same cell.' In fact the name is often written in that way ; thus, %y^ H^ /C. E. CURVISTELLATA. 217 EUPLECTELLA CURVISTELLATA nov. sp. Euplectella, Takeshita 19'. In the July number, 1900, of tîie Zoological Magazine (published by the Tokyo Zoological Society) appears a brief notice by Mr. Takeshita, of the Kagoshima Middle School, of the discovery of Eupledella sp. off the southern coast of the Province of Satsraa in Kyüshü, A specimen was obtained from a fisherman in that district, where, it is said, Euplectelhie are often brought up from a depth of 70-100 hiro (say, 100-142 m.) by the hooks of long-lines, sometimes to the number of three or four at a single haul. At my request, the specimen was kindly sent to me for examination, just in time to insert ils description in this work. It was in a badly dilapidated condition but nearly entire, the parts being sufficiently preserved to give a fairly good idea of its original appearance. It may at once be stated that it most closely resembles E. oweni, but differs from that species in its peculiarly characterized oxyhexasters, which seem to sufficiently justify its erection into a new and distinct species. I propose to call it E. Gurvistellata. However, with more materials at hand, it may possibly turn out advisable to regard it as only a variety of E. oweni. The body is 165 mm. long. On restoring it from a collapsed state to a tubular form it is found to be only perceptibly bent and to be slightly ventricose in the lower half. The greatest breadth measures 37 mm., against 30 mm. in the region of the cuff. The manner of juncture of the body with the sausage-shaped 218 I. IJIMA : HEXACTINELLIDA. I. mass of the basal tuft is exactly the same as in E. marshalli. The bulky basal mass, 115 mm. long, includes sand, pebbles and fragments of shells, etc., among its fibers, indicating the character of the sea-bottom. The external surface, though much damaged, may safely be said to be tolerably even. Parietal ledges, if at all recognizably developed, must have been rather insignificant and of only oc- casional occurrence. A small portion of the cuff, 2'/2 mm. in width, remained to the specimen. In all these respects and in the aj^pearance of the sieve- plate, the resemblance to E. oweni must be said to be very close indeed. The same is true of the essential points in the spiculation. Of the sheletal framework I have counted 40 circular and 36 longitudinal beams. The principalia to both these beams are large oxystauractius. In the sieve- plate beams, they are mostly represented by oxydiactins. Let it also be expressly mentioned that the oscularia are mainly compass-needle-like diactins as in E. oweni. The basal a7ichoring spicules likewise exactl}'' as in that species. The dermalia of the usual shape have the distal hilt-ray 00-160/^ long and G-O'/o/-« broad near the spicular center. The floricornes measure 1)1 ii in average diameter. GrapMo- comes in intact condition have not been found ; but their presence is not to be doubted, since the rhaphidial terminals occur here and there near the external surface in the usual disposition, though not in great numbers. Now what constitutes the characteristic feature of this species is the somewhat unique appearance presented by the oxyhexasiers. These are very abundant everywhere in the wall. Compared with those of E. oweni, they are decidedly larger, E. CUKVISTELLATA. 219 measuring 7Ô-100// — on an average 90/^ — in diameter (against 50-70// of E. owenl). The rays, both principal and terminal, are somewhat stouter. (Compare the annexed wood- cuts with PL VI, figs. Ö & 6). More- over, the smooth and finely attenuated terminals, of which there are 3 or 4 (sometimes 5) to each short principal, are near their free ends ahvays more or less distinctly curved, frequently in an almost hook-like manner. For the rest of their length, the terminals are nearly straight. The bending takes place apparently without any definite rule as to its direc- tion. The terminals belonmns: to the same principal are sometimes bent all alike outwards, i. e., away from the axis of the principal. At other times they may be bent some outwards and some inwards, or in any intermediate direction. Two oxyliexasters from E. curvistellata. Magnified 440 X. Finally, let it be mentioned that the specimen contained in its gastral cavity a pair of Spongicola venusla, known also to in- habit E. aspergillum, E. marshalli and E. oiceni. Besides, I have a specimen of Ilyalonema sieboldi harboring a pair of the same Crustacea. The identity of the inmate may be taken as suggestive of similar bathymetrical and other conditions under which the above-mentioned Hexactinellid species live. 220 I. IJIMA : HEXACTINELLIDA. I. Regadrella O. Schm. The genus Regadrella bad long been known in a single species, R. phœnix, wbicb was first described by O. Schmidt ('8o). In '96 I referred to tbat genus a species wbicb I briefly described under the name of R. oh'inomana. Recently F. E. Schulze (19') bas described bis R. decora, respecting wbicb I greatl}' doubt if it can be held to be distinct from my R. ohinoseana. In tbe present Contribu- tion will be added anotber new species to be called R. homeyamai, wbicb shows an indubitably close affinity to R. phœnix O. Schm. Perhaps we have still anotber species in tbe specimen wbicb will also be described later on, provisionally identified as R. phœnix. Practically, three species come into question in determining the generic status of Regadrella ; viz., R. phœnix, komeyamai, and ohinoseana. Granting tbat all these were correctly referred to one and the same genus, the generic diagnosis will have to be drawn up somewhat as follows : Tubular or saccular Euplectellid, firmly attached to the solid substratum by a hard, knobby base. Su- perior end having a sieve-plate, which may be re- presented by remnants of its beams — a number of spicular rays in a wreath. Lateral wall with round parietal oscula. Skeletal beams running obliquely; with strong diactins as their principalia; fused together in tbe lower part of tbe body. Par en c by m alia accessoria thin-rayed diactins, hexactins, &c. Hexasters of 0 kinds: 1) floricome, 2) graphiocome and 3) either onychaster or oxyhexaster, w^hicli latter is generally reduced to the form of oxystauraster. EEGADRELLA. 221 To give the differential character of each species in the form of a key : «. — Parietal ledge only indicated. With on5'cliaster (no oxyhexaster or oxystauraster). Large oxypentactin parencliyiualia present along the superior rim of the lateral wall. Pricl^ly parenchymal oxyhexactins not pres?nt. a'. Witli true sieve-plate. CufF rudimentary. Without pro~tal needles in tufts on the lateral wall li. fliœnix O. ScHM. 6'. Sieve-plate represented by a spicular wreath (corona) guarding the superior terminal osculum. Cuft' well developed. With long prostal needles in tufts on the lateral wall R. komeyamai Ij. b — Parietal! ledge conspicuously developed. Without large oxypentactin parenchymalia along the superior rim of the lateral wall. With oxyhexaster, predominantly in the form of oxystauraster. Numerous small prickly oxyhexactins present in the parenchymalia. (True sieve-plate present, surrounded by a well developed cuff. Without tufts of prostal needles) R. okinoseana Ij. The idea of removing E. okinoseana altogether from the genus has often suggested itself to my mind. In fact I think this step might be taken with some practical advantage to the systematic. The presence of oxyhexasters and oxystaurasters instead of onychasters, and also of small, prickly or spinose, parenchymal oxyhexactins in large numbers, keep this species somewhat apart from the other two, which iîiter se show an essential agreement in spiculation. Another not unimportant dis- tinction from those species seems to lie in the fact that in it the large oxypentactin parenchymalia, which in E. phœnix give a strong support to the sieve-plate and in B. homey amal supply the coronal rays, are wanting. Perhaps it may not be altogether inappropriate to associate R. okinoseana generically with Corbitella sjyeciosa { = Habrodidyum speciosum Q[]0Y & G KiM.k^J)), whicb, to judge from W. Thomson's ('68) statements, seems to agree in a measure with that species, amongst other points in being in possession of oxyhexasters and evidently also of small smooth- rayed hexactins which may correspond to the spinose parenchy- mal oxyhexactin of R. okinoseana. Howver, in view of uncertain- 222 I. IJIMA : HEXACTINELLIDA. I. ties in our knowledge of Corhitella, R. ohinoseana may after all for the present best be left as it is. Regadvella is evidently mucli more nearly related to Tccgeria F. E. ScH. than may appear at first sight. The coronal wreath of T. jndchra, the only known species of that genus, is to be considered as of only specific rather than generic value, as will be enunciated anon under R. komeyamai which possesses the same structure. The spiculation in the two genera is to a far- reaching extent, essentially similar. The small, spinose, paren- chymal oxyhexactin of R. ohinoseana is represented in T. pulchra Chall.-Eep., pi. xr, fig. 2). All the three kinds of hexasters seen in both R. phœnix and R. Icomeyamal are here likewise present. Floricomes and graphioconies were mentioned as such by F. E. Schulze (/. c, p. 95) in T. pulchra ; for onychasters I take the hexaster-form which that writer has specially described as having 4-6 small hooks j^i'ojectiny transversely at the extremity of rather slender terminals. Now, what constitutes the most characteristic feature of T. pulchra is the presence, in addition to above-mentioned hexaster-forms, of well-developed discohexas- ters, whose arched terminal disc bears six strong hooks. The spicule called by F. E. Schulze the * discohexact ' (/. c, pi. xr, fig. 3) is, in my opinion, to be classed under the above dis- cohexaster simply as a case of hexactiuose discohexaster.'" Since now such a discohexaster differs from an onychaster merely in the more strongly developed state of the terminal disc or whorl of teeth, the distinction of Tœgcria from Regadrella may be said to rest on nothing more than the relative degree of the develop- ment of parts in certain discohexasters. *Tlie hexactiuose discohexaster apparently occurs also n\ Eudidyum elcgans desciùbed by ]\Iaksiiall ('75)- With a better knowledge, than we at present have of this species, it may possibly be found necessary to icgard Tcc(/aia F. E. Hvn. as only a synonym of Eudktyiim MaRöii. K. OKTNOSEANA. 223 REGADRELLA OKINOSEANA Ij. Pis. VII and VII r. B. oicinoseana, Ijima 'q6, p. 250. E. decora, F. E. Schulze ig', pp. 30-34, 43 ; pl. vi, %s. 10-18. From time to time several specimens of this exquisite species have been obtained by Kuma in the Sagami Sea, though mostly in fragments. The exact localities are as follows : Gokeba, about 572 m. (400 ÄzVo = 313 fms.). Okinose, about 358 m. (250 h{ro = 19Cj fms.). Outside Okinose, over 500 m. (350 hi, '0 = 274: fms.). Inside Okinose by Ena-line, between 429 and 572 m. (3-400 Airo = 235-313 fms.). Ike-line"'"' by Mocliiyama-line, about 832 m. (580 hiro = 454 fms.). Many of the specimens bear at the base a sample of the bottom, invariably a tufaceous rock, to which they are firmly attached. If I am right in regarding II. decora F. E. Sch. as identical with it. okinoseana, a very wide distribution is to be ascribed to the species. Schulze's type of 7t. decora came from a spot in the Indian Ocean, SW. of Cape Comorin and 787 m. deep. Attached to the skeletal stumps of dead specimens from Okinose Inside, I have found an interesting series of the young, *Tlus line, not given in PI. XIV, lies a short distance to the east of the Jögashima Lt. house line. 224 I. IJIMA : HEXACTINELLIDA. I. which will be described after I shall have finished with my accounts of the full-grown specimens. General Characters of Full-Grown Specimens. The species may attain a very considerable size. A superb specimen was that which was purchased by Prof. A. Ac4ASSiz in Yokohama and taken to the United States. The vase-like body measured 400 mm. in height. Diameter of the sieve-plate 140 mm. Width of the cuff 60 mm. in the broadest part. Some of the parietal ledges as high as 55 mm. Another large specimen, belonging to IMr. Alan Owston, measured 420 mm. in total height. The upper portion of the body was rather abnormally inflated into a bulbous shape, presenting diameters of 270-330 mm. Near the basal end the breadth measured 30-33 mm. In the Sei. Coll. Museum there is one exquisite specimen which is not very large but is preserved in alcohol in an almost perfect condition. This is shown in PI. YII, fig. 1, in half natural size, and will here be described somewhat in detail. It is of an elongate vase-like shape, 185 mm. high, and of very irregidarly corrugated external aspect on account of the parietal ledges. Tlie breadth measures 72-80 mm. across the cuffed upper end ; 40-58 mm. directly behind the cuff; 6f5-75 mm. at the middle of the body ; and 15-20 mm. at the contracted base close to the solid, irregularly lobed basal mass. The cross- sections of the body-wall near the upper and the lower ends are nearly oval ; it is more irregularly shaped in the middle. The sieve-iolate is well arched and oval in outline, measuring E. OKINOSEANA. — GEN. CHAEACTEKS. 225 37 mm. by 54 mm. in diameter. In general appearance it agrees well with Eiiplectella marshalli. The beams are mostly y^-l mm. wide, the thinner ones being somewhat laterally compressed but the thicker ones so flattened as to present broader sides exter- nally and internally. In two or three places they join together to form nodal plates 2-4 mm. in width. Among the beams no radial spoke-like arrangement can be discerned. This is in accordance with the absence, in the sieve-plate border, of such large oxypentactins as are present in other species of the genus, w^hich might give rise to a radial arrangement of the beams. The meshes are oval, oblong or angular but always with rounded corners. They measure more usually 1-472 mm. across. An interesting fact is that most of the meshes are each provided, like the paiietal oscula, with a thin, narrow, iris-like membrane that leaves a round aperture in the middle. The cuff is very broad with a wavy edge-line, measuring in places 22 mm. in width, and is irregularly undulating, being- directed obliquely upwards and outwards. It is soft and moderately thick. The free edge bears an inconspicuous row of projecting spicules, not more than half a millimeter in length. The ledges are low in a narrow zone directly behind the cuff; so also in the basal section of the sponge. In the remain- ing major part of the wall, they are very prominently developed in the form of thick, round-edged and extremely irregular ridges, which vary greatly in height at different points. They may in places be 20 mm. high and 4 mm. or more thick. They frequently branch in their course, making it difficult to determine the general direction they take. Here and there are seen evident 226 I. IJIMA : HEXACTINELLTDA. I. signs of the ledges liaviiig fused together secondarily, occasionally leaving an arch-like or tunnel-like passage underneath. Some- times the elevation surrounds a depression containing a solitary parietal osculum and at other times incloses an irregular valley- like space in which several oscula may lie side hy side. Like the cuff", the ledges are soft and can be easily torn away from the sponge-wall, except in the basal region of the body where they are firm owing to the extensive amalgamation of the mega- scleric elements. The parietal oscula are round, up to 3 mm. in diameter, and are surrounded by a narrow oscular membrane as in Eupleclclla. Their distribution must be said to be irregular, being situated 3-8 mm. and sometimes even 15 mm. distant from one another. The surface of the parietal ledges presents for the most part a rather close-grained texture. However, towards the base of the ledges and over the depressed area around the parietal oscula, there are visible, by the aid of a hand-lens, the usual dermal latticework of a most delicate nature, extending itself close to the oscular edge. Beneath this layer are discernible the variously sized apertures of incurrent canals, measuring up to about 1 mm. across. The same apertures are also exhibited by both the superior and the inferior surfaces of the cuff. Leaving the ledges out of consideration, the sponge-wall must in general be said to be thin, except at the much thickened, blindly closed end at the extreme base. In most places the wall does not exceed 2'/._. mm. in thickness. Nevertheless, the entire E. OKINOSEANA. — GEN. CHARACTERS. 227 specimen is sufficiently rigid to keep its shape when taken out of the spirit in which it is preserved. To illustrate the appearance of the parietes on the inner side, may serve PL YII, fig. 2. It is taken from a specimen wanting the upper end but otherwise well preserved and which has been longitudinally bisected for this special purpose. The two kinds of openings visible on this side present much the same appearance as in Euj)lectella. The openings of the excurrent canals usually measure less than ly. mm., but occasionally 2 mm. across. Unlike in Euplectella, the narrow ridges produced by the main skeletal beams are all more or less obliquely disposed and intersect one another at various angles. However, it can be distinctly observed in this as well as in other specimens that certain beams or spicular bundles, lying innermost in the wall and evidently corresponding to the circular beams of Euplectella, are relatively more transversely disposed than others situated nearer the exterior. As could be observed in the complete specimen before de- scribed, the skeletal beams of the parietal wall, at the upper end, pass directly into those of the sieve-plate. If, therefore, the cuff and all other loose parts be made to fall off the megascleric beams, the framework of the lateral wall should be seen to con- tinue itself without any demarcation into the sieve-plate, much in the same manner, I should think, as in that old specimen well-known as the type of Corhitella speciosa (Q. & G.). The lower end of the sponge-body shows the larger paren- chymal spicules and their bundles firml}^ ankylosed by synapticular fusion, which may extend above for about one half or more of 228 I. IJIMA : HEXACTINELLIDA. I. the entire body-length. So that, after death and the washing away of all the loose spicules, only the inferior portion of the sponge remains with any degree of persistence as a perforated but compact cup with a solid knobby base. In this condition are several specimens now before me. Fortunately, all these, dead stumps as some of them are, still contain the spicules charac- teristic of the species, which puts the identification beyond the reach of doubt. PI. VII, fig. 4 represents, in natural size, the macerated remnant of the skeleton of a comparatively small and young specimen. It consists for the most part of fused spicules. I may remark that the general appearance of this specimen strongly reminds me of one of the two specimens on which O. Schmidt ('8o, p. 4G ; Taf. Vir, 3 a) based his Rhahdodictyuiii delicatum. PL VII, fig. 3 shows in a typical way the dead skeleton of a large specimen. The wall exhibits externally an irregular network of hard and more or less prominent ridges. It scarcely needs to be mentioned that these arose by the soldering together of parenchymal bundles in the parietal ledges. In the depressed spaces bounded by the ridges are situated single, less frequently several, roundish gaps, indicating the position of parietal oscula in the living state. Immediately around the gap, the wall forms a netted plate made up of a number of spicular strands branching oil* from neighboring coarser bundles and running tangentially in all directions. The coarser bundles, some of which may be nearly 1 mm. thick, are seen to run in the main in two opposite- ly directed, oblique sets. In their course they iTcely split, unite and intersect or pass through one another, thus giving rise to an irregular basket-work which may, on that account, be readily distinguished from the more regularly framed skeleton of Uic- K. OKINOSEANA. — SPICULATION. 229 plectella. On the gastral surface there are seen the buDclIes already referred to, the course of which more closely approaches the transverse than that of any of those visible on the external side. — Superiorly, the fusion of spicules gradually diminishes in degree and extent; the interweaving of the fused bundles becomes looser ; finally, each of these runs out into fretted, tuft-like strands. At the same time the hard external ridges disappear, becoming replaced in the living state by the loosely supported ledges, which are of course lost after maceration. Spiculation. The two specimens shown in PI. YII, figs. 1 & 2 (Sc. Coll. Mus. Nos. 487 & 488), were principally made use of in my study of the spiculation in full-grown individuals. The principalia in the parenchymal bundles are large oxy- diactins, which may attain a length of 35 mm. or more and a breadth of 220 !'■ in the thickest portion. They are nearly straight or gently bent, without an elbow-like bending at the middle. Towards both extremities they attenuate to thin, smooth or rough-surfaced, pointed ends. — The coiidtalia, accompanying the above principalia in a copious quantity, are mainly slender diactins, — not thetactins as in EuplecteUa. They are usually 10-16 IJ- thick ; generally smooth but occasionally annulated or tubercled at the spicular center ; subter min ally rough-surfaced., the very end being smooth and rounded or conically pointed in the usual way. Just the same diactins as the comitalia occur in all parts of the wall either in loose arrangement or in strands by themselves. 230 I. IJIMA : HEXACTINELLIDA. I. Among tlie pareuchymalia there also occur not infrequently smooth oxyhexaciins of comparatively large or medium size. The rays are subterminally rough-surfaced ; up to about 17 ii iu thickness near the base ; varying in length not only in different spicules but frequently also in the same spicule. Such a paren- chymal oxyhexactin frequently occurs in the choanosome without apparent definiteness as to the orientation of its rays in relation to other skeletal parts. Sometimes however, there have been found some whose size, shaj^e and situation, suggest that they are reserves, as it were, of certain dermalia. They seemed to require only to be pushed out more or less, in order to be classed wntli the dermalia. On the other hand, there are occasionally found similar oxyhexactins participating with one of the axes in the formation of a parenchymal strand. The said axis may then be greatly prolonged in excess over the other two. I think I may say that the oxyhexactins represent an intermediary between the dermalia and the diactins which make up the main contingent of the parenchymalia. Characteristic of the species is the abundant occurrence of a kind of intermedial parenchymal oxyhexactins, which, for the sake of reference, may be called the in'icroxyJicxaciLU>< (PI. VIII, figs. 24-26, 32), This is exceedingly variable in size but on the whole it is small, usually measuring 175-300/^, sometimes only 110/''-, in axial length. The rays are 4-8//, rarely as much as 20 1^, thick at base ; straight ; attenuating to a fine point. They are invariably characterized by having the entire surface beset with numerous, vertically out-standing, minute prickles. The 2)rickles are more pronounced in some cases than in others and are decidedly spiny. The axial filament in each ray reaches E. OKINOSEANA. — SPICULATION. 231 right up to the pointed end, placing it beyond the reach of doubt that we have Iiere to do with a true hexactin. Spicules of similar or exactly the same appearance are known from Tœgeria as well as from WaUeria. The microxyhexactins are present in all parts of the paren- chyma. They seem to be more abundant in the inner than in the outer trabecular layer. They mostly occur loosely, sometimes tightly clasped in the bundles of parenchymal diactins. Further, I have seen them situated and arranged after the manner of canalaria in places in the excurrent canals. In the deeper parts of the parietes I have occasionally met with isolated spicules, which somewhat differed from, but seemed to integrade with, the microxyhexactins. We here liave to deal with rather small pentactins or hexactins or such hexactins as approach a pentactin by the reduction to a greater or less extent of one of the rays (PI. VIIT, figs, 27, 28, 33). The rays differ from those of microxyhexactins in being somewhat thicker and less tapering, in having rounded or bluntly pointed ends and in being sparingly supplied with prickles only near the end. The prickles, however, have sometimes been found to extend nearly all over the rays, though in a weak state of development. I am inclined to regard these spicules as representing parenchymal microxyhexac- tins in the way of ditferentiation towards gastralia or canalaria. The hard basal mass consists of a rigid, close and irregu- larly meshed framework of siliceous beams, which bear on their surface sparingly and unevenly distributed microtubercles. The beams arise by extensive synapticular fusion of all the paren- chymalia in this region, except the intermedial microxyhexactins, which, together with the hexasters, usually remain free in the 232 I. IJIMA : HEXACTINELLIDA. I. meslies. In addition to the same megascleric elements as are found in the upper part of the skeleton, there are contained in the basal mass a large quantity of peculiar hexactins, which occur nowhere else and which I have called the hasidictyonalia. The hexactins in question are comparatively small in size but have thick, plump- looking rays, which are nearly smooth or show a few microtubercles near their rounded ends (PI. YIII, fig. 34). The basidict3'onalia are at first loose but soon become soldered to the general framework of the region. PI. X, fig. 17, representing a small piece of the basal mass taken from 11. homeyamal, may just as well pass for the same of the present species ; in it some of the beam nodes are plainly the center of basidictyonal hexactins. The secondary deposit of siliceous matter over the surface and the synapticula3 irregularly proceeding from it often make the hexactins unre- cognizable as such externally, but the characteristic triaxial central filaments remain in the beams. The basidictyonalia seem to be of quite general occurrence among those Lyssacine Hexactinellids which are attached to hard foreign bodies directly by a part of the wall, and whose sj^icules undergo extensive ankylosis in the basal region. F. E. Schulze figured them from the firm stalk of Cratcromorpha meyerl (Chall. Kep,, pi. Lxr, figs. 5 & G). To them I refer also the rigid reticulum of spicules described by the same writer ('gg, p. G4) from the buds of Rhahdocalyplus mirabilis, which fact I have al- ready had occasion to mention on p. 186 (foot-note). And, I shall have to demonstrate their presence in a series of other forms in the course of these Contributions. The framework of the basal mass is especially close meshed, on account of an excessively abundant development of synapticular formations, in the bounding surface which is in direct contact R. OKINOSEATsA. — SPICULATION. 233 with the solid substratum. Here the meshes are not wider than the beams themselves. The irregular cribellate plate thus formed was known to F. E. Schulze ('87, p. 3S ; pi. LXiv, fig. 3) from Bhahdocalyptus inoUh, etc. Spicules whicli might correspond to the o^cularia of Euplec- tclla were not noticed in any of the specimens, except in a rather small individual from Okinosé (Sc. Coll. Mus. No. 490). In this, the iris-like membrane of parietal oscula (PI. VIII, fig. 37) was supplied with an abundance of small spinose hexactins, most of which differed from ordinär}^ microxyhexactins in having somewhat thicker rays terminating in rounded ends, and also in frequently having one or more of the rays reduced in length. Thus, they were not uncommonly pentactins, and occasionally even diactins, in all of which the suppressed rays were re- presented by knobs or rudiments of variable length. The presence of transitional forms, however, clearly indicated their derivation from microxyhexactins by modification. They are evidently a sort of spicules which is of inconstant occurrence in the species. The dermalia (PI. YIII, figs. 14-18) are hexactins of vari- able size and strength. jMany of them may be said to be sword- shaped with the proximal ray more than twice as long as the paratangential rays ; while others, especially those of weak development, may have that ray of nearly equal length, or even somewhat shorter than these. The distal ray is always distinguishable b}^ its comparative shortness, by its rounded or conical end, and by the relatively more numerous and more pronounced development of the microtubercles on its surface. 234 I- IJIMA : HEXACïINELLTDA. I. Frequently, but not always, the distal ray is slightly swollen near the end, presenting a club-like shape (fig. 18). All the other rays are tapering toward the pointed end and nearly smooth all over or snbterminally obsoletely tubercled. — The distal ray is usually 80-150 n long ; sometimes as short as to measure only 50 /i in length. The paratangential rays are 175-275 mm. long. In the strength of the rays, as also in the manner of arrange- ment of the dermalia, there obtains a noteworthy difference in different parts of the sponge surface. Slender-rayed are the dermalia in the depressed areas around the parietal oscula (figs. 17, 18). The rays measure about 6^/2 IJ- ill average breadth near the central node. The paratangentials are arranged so as to form a tolerably regular, quadrate meshed latticework, with meshes 150-300 n in length of sides. Whereas, on the parietal ledges, there occur on the whole somewhat larger and much stouter dermalia (figs. 14-10, 29), though these are by no means of uniform strength. On an average, the rays are here about 12 // thick near the spicular center. The larger dermalia seemed to increase in number as they approach the edge of the ledges, although even in this part there may occasional!}^ be intermingled such as are as weakly developed as any in the entire dermal system. Moreover, the dermalia occur on the ledges irregularly crowded, so that a re- gularly meshed latticework is not brought into formation. The greatest development is attained by the dermalia along the free edge of the cuff (Pi. VIII, fig. 13). They may not inappropriately be called the prostalia marginalia. The shape is sword-like. Total length up to 2\l.2 im«. The rays reach up to 45 f^ in breadth near base ; all of them taper towards the pointed end. The free distal ray may be 800 /^ long ; it is beset for the K. OKINOSEANA. — SPICULATION. 235 greater part of its length with erectly out-standing microtubercles. The paratangentials are comparatively very short (up to 240 ;« in length) ; they are rough-surfaced only near the end. The jDrolonged proximal ray is nearly smooth all over. An idea of the large size attained by the dermal hexactins on the cuff- edge, and of the variability in size of the dermalia in general, may be obtained by comparing fig. 13 with figs. 14-18, all of which figures are drawn on the same scale of magnification. Not that all the hexactins on the cuff- edge are uniformly large, but there are mixed with them smaller ones which connect them with the dermalia of the general surface. The distal rays, in forming the inconspicuous marginal row before mentioned, are accompanied by a number of rhaphides, not with diactin comi- talia. Noteworthy is the fact that along the cuff-edge as well as in certain parts of the ledges, the stronger-rayed dermal hexac- tins apparently take their origin among the parenchymalia and are subsequently added to the dermal layer from below. The gastralia are pentactins of various sizes and of irregular appearance in so far as the paratangentials are often not straight, and are of unequal length in the same spicule. The rays are frequently only about 175 /^- long, while at other times they are fully four times as long, with an average breadth of 17 /^. The unpaired distal ray may be shorter or longer than the average length of the paratangentials. The rudiment of a sixth ray is generally present in the form of a hemispherical knob. All the developed rays are subterminally faintly rough, the very ends being rounded or pointed. The gastralia are found in irregularly scattered distribution. 236 I. IJIMA : HEXACTINELLIDA. I. The paratangentials lie in direct contact with the parenchymalia and often run in association with bundles of these. It may be worth while to note that not infrequently some parenchymal fibers or strands intersect the paratangentials on the inner side. Rarely and exceptionally there occur stauractins or thetac- tins in the place of pentactin gastralia. Pentactin canalarla have not been observed. The floricomes (PI. VIII, fig. 23) are of typical form, meas- uring 98-107 /^ in diameter. The number of terminals in a perianth is usually 6, sometimes 7 or 8. The terminal disc bears rather strong teeth, as a rule 3 (seldom only 2) in number. The inner border of the disc, when seen in lateral view, is indicated by a rounded angular bending of the contour-line on that side. The floricomes are very common in depressed and therefore protected positions of the external surface. In regions immedi- ately around the parietal oscula, every distal ray of the dermalia may be said to bear a floricome on its tip. In exposed parts of the ledges they occur but rarely, whether in the position just mentioned or in the subdermal region. Graphioco/iies in an intact state are exceedingly rare. Of common occurrence is their central portion (PL VIII, fig. 36) after the loss of the rhaphides. Such a relic consists of six principals, about 3 y- thick and 15 /^ long, each bearing at its ..ijßud a small disc, the outer surface of which is beset with short basal remnants of the rhaphides. ^The rhaphides, 180-200 !'■ long, occur very abundantly, cither in shelves or in a scattered state, in the ectosomal region. As E. OKINOSEANA. — SPICULATION. 237 a rule their one end is outwardly directed, and often freel}'' projects more or less beyond the external surface. On several occasions complete grapliiocomes have been ob- served with terminals measuring only 30 /^ or 20 i'- in length. These were undoubtedly in immature stages of their development. PI. VII, fig. 9 represents one such developing graphiocome taken from a young specimen. Although the floricome and the graphiocome must be said to belong par excellence to the external trabecular layer, yet certain observations seem to prove that both may sometimes arise in the inner trabecular layer as well. In the latter layer there have at times been found floricomes apparently young in ap- pearance, and that too under circumstances which made me disinclined to assume that they came there by dislocation. As to the graphiocome, a small and young specimen of the species showed several developmental stages of that hexaster, by the side of the relics of old ones, inside the chamber-layer and close to the gastral surface. True oxyhe.msters (PI. VIII, figs. 19, 20) occur only occa- sionally and may therefore be easily overlooked unless a special search be made for them. On the other hand, their derivative, the o.vijstau raster (PI. VIII, figs. 21, 22, 3ö), is abundantly present in both the outer and the inner trabecular layer, perhaps somewhat more numerously in the former than in the latter. In both kinds of the oxyasters, the diameter usually measures 68-100 /^-, exceptionally only about 50/^-. The smaller sizes refer as a rule to oxystaurasters, while the largest size is found espe- cially among the oxyhexasters. The principals are of moderate length and relatively slender, being about 11 y- long as measured 238 I. IJIMA : HEXACTINELLIDA. I. from the central point of the axial cross and about o'/i !'■ broad on an average. The slightly swollen end of the principals bears »3-5, rarely only 2 or more than 5, terminals in a diverging tuft. In the case of oxyhexasters, the number of terminals to each principal frequently runs up to 8, 10 or even more. The terminals are nearly thrice as long as the principal ; they are smooth, tapering, generally not straight but bent in a somewhat wavy manner. When numerous terminals form a tuft, they do not arise in a regular circle, but one or more may occupy a more central position than the rest. The tuft may be so diver- gent that any two opj)osite standing, outermost terminals form an angle greater than 90°. F. E. Schulze has called attention to the fact that in the oxystauraster in the type of his II. decora, those terminals lying in the plane vertical to that of the four principals stood out from the axis much more divergently than any other terminal. Something like this has also been noticed by me in certain instances, but not with any such degree of constancy as justifies one in deducing a rule therefrom. The oxystauraster is undoubtedly derived from the oxyhex- aster by the suppression of one of the axes. I have once seen a form with live principals and as many tufts of terminals. At the time I thought it was a genuine oxypentaster, but when I afterwards wanted to confirm my impression that the absence of the sixth arm was not due to a mechanical breaking off, I unfortunately failed to rediscover the rosette. In the few cases of oxystaurasters, in which I have specially entered into the examination of the axial filaments, I have seen no trace what- ever of a third axial filament. R. OKINOSEANA. — YOUXG SPECIMENS. 239 Respecting the spiciilation of the sieve-plate, I have to notice the following : The main support of tlie beams is afforded by spicular bundles whose components are essentially the same as in the parenchymal strands of the lateral wall. The only point of difference seems to consist in the fact that many of the diactins in the sieve-plate beam are of unusual shortness. These may be called compass-needle-like, with or without knobs at the middle. In extreme cases they are so short as to be only 250 ,« long, with a breadth of about 25 />« near the middle. Spinose microxyhexactins occur but rarely. Floricomes and oxyasters have not been found, but the sheaves of graphiocome-terminals are common. The dermal hexaetins, which occur very closely crowded on the external side of sieve-plate beams, deserve special mention (PL VIII, fig. 31). The rays are thick and short, measuring 80-lC)0 IJ- in length and 15-27 /^ in thickness at their base. All the six rays in the spicule are nearly equal in length and in gene- ral appearance. They are generally tapering, minutely tubercled on the outer part, and end either rounded or in a j^oint. On the inside of the beams occur similar spicules which are however mostly pentactins but occasionally stauractins, and which are undoubtedly to be regarded as gastralia. The unpaired ray of the pentactins dips into the parenchymal bundle. The said spicules are present at wide intervals, so that the parenchymal bundles are largely exposed to view^ on this side of the sieve- plate. Young Specimens. The Science College collection contains an interesting series 240 I. IJIMA : HEXACTINELLIDA. I. of young specimens which I refer to the present species, not- withstanding certain discrepancies between them and the adult forms in both nnicroscopical and microscopical respects. The series comprises different stages, from one in which the body is smaller than a grain of rice to such as have the characteristics of the adult nearly completely developed. The specimens will here be described in the order of their development, beginning with the youngest. 1. On two stumps of dead skeletons (Mus. Xos. 309 & 370), which I have identified as H. okinoseana from microscopical examination of their spicules and both of which were obtained from Inside Okinose during January, 1895, were found several small and delicate Hexactinellids of a club-like or elongate ovoid shape. They were attached by a small basal expansion at the narrowed lower end to the beams, on both the inside and out- side of the dead specimen. The smallest individual of the lot was only 4 mm. long with a breadth of about 2 mm.; the largest, 13 mm. by G mm. In PI. YII, fig. o are shown three of the small specimens in question, magnified about V/._> times. The rounded upper end always shows a simple, round or oval oseulum which leads into a deep, tubular, gastral cavity. The dermal surface is smooth and nowhere interrupted by parietal gaps. Thus, in macroscopic respects, the specimens can scarcely be distinguished from those of either Vitrollula or Leucopsacin^ ; and even after gaining a knowledge of their spiculation, I was at first far from recognizing them to be the young oî E. okinoseana, the same in species as the dead sponge to which they were attached. In the first place, the dermalia (PI. VII, figs. 8, 10, 11) are E. OKINOSEANA. — YOUNG SPECIMENS. 241 exclusively pentactins, not liexactins as in all mature Euplectellids. The paratangential cross measures 275-650 /^ in axial length and is generally slightly arched in conformity with the curvature of the surface (PI. VII, fig. 8). Its rays are tapering and termi- nate in pointed or conical! y obtuse ends. Besides the usual roughness of surface near the end, they show a number of ob- solete tubercles throughout their entire length, except along their inner side where the tubercles are nearly or quite absent. Generally but not always, the center of the paratangential cross exhibits on the outer side a gentle swelling. The unpaired prox- imal ray is developed to a length that usually exceeds by twice or even thrice that of the paratangentials. Like these, it is obsoletely rough ; but the roughness gradually loses itself proximad towards the finely pointed end of the ray. — Seen on the dermal surface, the paratangential crosses are rather irregularly disposed to one another or show a tendency to arrange themselves into a quadrate-meshed latticework (PI. VII, fig. 10). Sections of the wall show that the elongated proximal ray reaches with its inner end nearly or quite to the gastral surface (PI. VII, fig. 11). Ko special gastralia are present. Along the gastral surface as well as in the deep part of the wall, there occur fine paren- chymal diactins, mostly arranged in obliquely running and intersecting strands. In some of the specimens was observed the spinose microxyhexactin in isolated occurrence, while in others this kind of intermedial spicule seemed to be as yet not at all developed. Much more constant and common is the grajDhiocome. De- tached sheaves of rhaphidial terminals, 170/^ and more in length, are to be seen in abundance in the periphery of the wall. The central relics of the graphiocome were also seen in 242 I. IJIMA : HEXACTINELLIDA. I. fair numbers. Intnct graphiocomes with terminals that had not 3^et reached their full-length (PL VII, fig. 9) were several times met with. No other hexaster-form has been found in these little specimens. What systematic position to assign to the specimens was at first a great puzzle. The presence of pentactin dermalia seemed to make against tlieir being regarded as Euplectellids, while the graphiocome pointed to their being at least a close ally of that family. Fortunately, however, I have found other young speci- mens which seem to represent transitional stages that lead over the simple spiculation of those little specimens into the more complicated system of the mature R. ohinoseana. 2. From still another skeletal stump of R. okinoseana (Mus. No. 41)0, from Inside Okinose, March 1898) was taken a young specimen of an elongate ovoid shape, 15 mm. long and 7-9 mm. broad. It is therefore considerably larger than the largest in the last described lot. A simple osculum, 2^2 mm. in diameter, is situated in the upper end. The dermal surface is no longer smooth but uneven. This is caused by the presence of small depressions, several of which have broken through the wall, while many others still remain closed. There can be no doubt that we have here to deal with the first formation of parietal oscula ; the mode of their origin is essentially the same as iu ^uplectella (p. 105). Examination of the spiculation also shoAved points of de- cided advance from the state noted in the last lot of specimens. The pentactin-dermalia present exactly the same characteristic features as in the latter. Only, mixed amongst them are oc- casionally found sword-shaped hexactin-dermalia, the distal hilt- K. OKINOSEANA. — YOUNG SPECIMENS. 243 ray of which is rough-surfaced and rounded at the end. The diactin-parenchyraalia and spinose microxyhexactins are present in greater abundance than before. Besides graphiocomes, there are now to be seen iloricomes, oxyhexasters and oxystaurasters, though as yet in quite limited numbers. In short the small specimen in question may be said to bear the essential characteristics of the adult Pv. okinoseana, except in the important respects that the terminal osculum is simple instead of being covered by a sieve-plate, and that the dermalia are predominantly pentactins. 3. By the side of the above specimen and on the same dead skeleton, was found another young specimen, which, though broken off in the upper part, must have had a somewhat larger body. A few parietal oscula open in the portion of the sponge-wall still remaining. As to the spiculation, the one important point in which this differs from the last specimen consists in the fact that the hexactin-dermalia are present in a notably increased number, — in about the same numerical proportion as the peutactin- dermalia (PI. VII, fig. 12). The hexactin-dermalia compare well with those of old specimens of Ji. okinoseana. I have noticed that the pentactins have on an average stronger rays than the hexactins and that the latter are very variable in size, the smallest having very slender rays indicative of its comparatively recent origin. Further, I may say that the paratangentials of the pentactins generally, though not always, overlie those of the hexactins. The evidences are in favor of the conclusion that, whereas the first formed dermalia are pentactins, those which begin to develop later are all hexactins, and that these are, as a general rule, added to the dermal layer from below. 244 I. IJIMA : HEXACTINELLIDA. I. 4. A still more advaoced stage of growth is represented by the two specimens figured in natural size in PL VII, figs. 6 & 7 (Mus. No. 461, from Outside Okinose, Dec. 15, 1898). One of these is of an irregularly tubercular shape and is 18 mm. high (fig. 6) ; it is attached to a piece of old basidictyonal mass, presumably of the same species. The other specimen (fig. 7) is a tubular sac broken off at the lower end. In both there is at the upper end a single, thin-edged and relatively large terminal osculum. In the uneven lateral wall, several small parietal oscula have opened, though many others are still represented merely by dimple-like depressions of the external surface. The im- mediate neighborhood of the terminal osculum is smooth-surfaced. The spiculation in both is essentially that of R. okinoseana ; but one point requires special mention, viz., that, though the der- malia are predominantly hexactins, there are still to be seen the original pentactin-dermalia in some numbers. In view of the fact that in the young of E. marshalli the delicate beams of the inceptional sieve-plate are exceedindgly liable to become lost by being broken off (p. 108), it might be questioned if a similar loss had not happened to the terminal osculum in the young specimens hitherto mentioned of R. okino- seana. Close inspection of the oscular edge in the two specimens just described, however, has seemed to show no sign of an un- natural severing off of any part of it. 5. A tubular specimen, contained in the same bottle as those referred to above under (2.) and (3.), has the loAver end wanting but must have originally measured at most 50 mm. in total height, with a diameter of about 12 mm. in the middle of the body. The general character of the external surface much E. OKINOSEANA. — YOUNG SPECIMENS. 245 resembles that of tlie young E. marshalli shown in PL IV, fig. 9 ; the ledges are indicated by low reticulate ridges with rounded edges, each depressed mesh containing a parietal osculum, of which there are many. At the upper end exists a transverse opening about 3 mm. wide. This is evidently the original single terminal osculum of the specimeu. Directly adjoining it on one side, but not on the other, is a small ill-defined area in which the thin sponge-wall is perforated by several, somewhat closely situated, irregularly angular gaps of various sizes. I take this area to be a beginning of the sieve-plate formation. However, I entertain some doubt as to whether I have seen it in quite its natural condition, since the tissues at the parts bore signs of laceration to a certain degree. Be that as it may, to my mind the first formation of the sieve-plate in Euplectellidœ probably takes place in the manner indicated. Thus, to the original single osculum at the upper end of the sponge, more oscula are afterwards gradually added in close proximity to it and to one another, converting the inter- vening part of the body-wall into the sieve-plate beams. The beams and nodes, after their establishment, may themselves also become perforated and thus may contribute to the multiplication of the sieve-plate meshes. 6. Finally I will mention a specimen 70 mm. liigli and 20 mm. broad at the widest part. It is attached to the basal ex- pansion of the same dead skeleton ou which the two young specimens mentioned under (2.) and (3.) were found. The body- form is peculiarly irregular, but this is undoubtedly due to accidental causes. The ledges are tolerably well developed. The upper end bears a definitely formed, but only slightly arched 246 I. IJIMA : HEXACTINELLIDA. I. sieve-plate of about 8 mm. diameter. The cuff is as yet narrow. Of the spiculation, I need mention only the fact that among the hexactin-dermalia are here and there intermingled the same pentactins, characterized by the peculiar distribution of micro- tubercles on tlie rays, as those which were found in the dermal layer of all the smaller specimens. The pentactin-dermalia were not noticed by me in the adults. With advance in age, they are either lost or become so outnumbered by the multitude of hexactin-forms as to easily elude being seen. From the data presented in the above, it may not be in- appropriate to draw up the following summarizations. Firstly as regards the macroscopic features : R. oHnoseana, in the first stage of postembryonal de- velopment, has a smooth and imperforate lateral wall, and is provided with a single terminal osculura. Soon (when the body measures, say, 18-15 mm, in length) the outer surface becomes uneven, and this stage is followed by the appear- ance of parietal oscula. Later (say, when over 20 mm. in length), the sieve-plate seems to become started by ad- dition of new terminal oscula to the one already present. The initial number of the meshes must at all events be very few. Individuals, say, 50-70 mm. long, show the cuff and ledges indistinctly developed. Secondly, with respect to spiculation : The initial dermalia are pentactins.''' To them, hexac- *Tlii.s remarkable fact is perhaps to be explained by assuming that at the time of tlie first appearance of tlie dermalia tlie superficial trabecular layer is of such limited thickness that it affords no space for the development of a sixth distal ray, the paratangentials being formed on its extreme outer surface. A i^henomenon in a measure analogous to this is seen in tlie development of Leucopsacus ortliodocus- In this si)ecies the first spicules formed in the E. OKINOSEANA. — SOFT PAETS. 247 tins are soon added, at about a period when the parietal oscula have begun to break through. Henceforth the derma- lia that newly arise seem to be all hexactins ; so that these soon greatly outnumber the original pentactins. The earliest formed hexaster is the graphiocome, the rhaphidial sheaves derived from it being common at the stage when the der- nialia consist as yet exclusively of pentactins. The formation of microxyhexactins, of the floricome, the oxy hexaster and the oxystauraster soon follows. Soft Paets. The trabecukc (PI. VIII, fig. 30) are abundantly developed in irregular cobweb-like arrangement. They are thin and filamentous, but here and there spread out into small film- like areas. The protoplasm is granular, moderately stained by borax- carmine. Nuclei belonging to it about 27, ,« in diameter, contain- ing a group of chromatin grains. Dermal membrane with meshes or pores of irregularly angular shape and of various sizes. Meshes separated from one another by thread-like, band-like or membrane-like trabeculîe. It is more or less extensively and continuously membranous towards, and on, the tent-like conuli produced by the distal rays of the dermalia. Archœocytes generally about 3 ji. in diameter ; found forming periphery of the larvae are stauraetins, which in the adults are replaced by pentactin-der- malia. — Probably the change in tlie dermalia, which I have endeavored to demonstrate in the young of B. okinoseana, is not peculiar to that species alone but is possibly common to a wide circle of forms belonging to the same family. I call attention to the stalked, evidently very young Hexactinellid figured by F. E. Schulze in the Chall. Eep., pi. XLii, figs. 5 & 6, and referred to by him as the ' undetermined Crateromorphid.' The essential agreement in spiculation between it and the little R. okinoseana I have desciihod on p. 240 {mh 1.) makes me believe that it is more probably a young Euplectellid, the pentactiu-dermalia of which assamably give place to hexactins in a later period of life. 248 I. IJIMA : HEXACTINELLTDA. I. in variable numbers small groups on the outer surface of chambers (PL VIII, fig. 29). Nuclei like those of the trabeculœ in both size and appearance ; with definitely circumscribed cell-body con- sisting of granular protoplasm. Tkcsocytes (PI. VIII, fig. 30 ; th.) found in some numbers scattered on both the outer and inner trabeculse ; 7-1-3// in dia- meter. The contents consist of réfringent, irregular granules or of variously sized spherules. In the same preparation, they are sometimes colored by borax-carmine and sometimes not. In the latter case, they present a yellowish-olive tint ; amongst them the nucleus may be discerned as an indistinct red spot. Chambers cup-like or thimble-like ; diameter 45-75 1'- (on an average 57/^). Choanocyte-nuclei small (about l7o/^ in dia.), pale, vesicular, without conspicuous chromatin grains in tlie in- terior ; flattened, when seen in profile ; generally ^-^ /^ apart from one another. Beams of the reticular membrane thin, granular ; meshes distinctly open (!). MlSCELLAîs^EOUS NoTES. F. E. Schulze (ig', p. 33), when he was describing his R. decora from small fragmentary pieces, was aware of the fact that that species might possibly prove to be identical with my R. olcinoseanaj which was known to him from the preliminary description 1 have given in the Zoologischer Anzeiger in 'g6. If the two species are identical, as I consider them to be, the fault which induced Schulze to create a synonym must be said to have lain chiefly in the brevity of that description of mine. Schulze was led to regard the two species as distinct, though very closely resembling each other, from a consideration of the R. OKINOSEANA. MISCELLANEOUS NOTES. 249 following three points in the spiculation of the tyjoe of R. decora : 1) The absence of oxyhexasters, which had been mentioned by me as present in the Japanese species ; 2) the occurrence of large and strong oxydiactin parenchymalia principalia (15-20 mm. long, 100-200/^ thick), of which I had made no mention; and 3) the f\ict that the outer radial ray (120 /-< long) of the dermalia was not particularly short, wliereas I had called it short in my sj)ecies. As regards the first point, it is to be remarked that the oxyhexaster and the oxystauraster, as they occur together in li. ohinoseana, are so closely similar, except of course in the point indicated by their names, that they scarcely deserve to be made into separate categories of much systematic significance. Moreover, the oxyhexaster occurs only occasionally and in numbers which, though subject to variation according to indivi- duals, may be said in general to be insignificant in proportion to those of the other oxyasters. It may therefore under certain circumstances be easily overlooked ; and besides, I think that in individual cases it may even be really entirely wanting, without, on that account alone, affecting the specific status. The second and the third point will have lost their weight as distinctive specific characters from what I have given in this Contribution for the size of the S2)icules in question from R. okinoseana. One point concerning the dermalia seems to require a remark. These were described and figured by Schulze as tolerably uniform in size and in the strength of the rays. This is also the case in my specimens so far as those on the flat or depressed areas of the sponge-surface are concerned; it is on the ledges, especially toward their edges, that the dermalia are sub- ject to a considerable variation in these respects. That Schulze 250 I. IJIMA : HEXACTINELLIDA. I. made no mention of this variability is probably due simply to the fact that the ledges were not represented in the pieces examined by him. After all I do not see in the organization of R. decora any tangible point by which it may be upheld as a valid species. Malformations in parts of the body, brought about after the healing of injuries received by the sponge-wall, are as common as, and present appearances similar to those in Eupleciella. Let mention be made here of a remarkable case of regeneration that came under my observation. Of a medium-sized specimen there remained only a small lateral piece of the wall at its base, standing on the basal ex- pansion ; the rest of tlie body had been torn off and lost. The remnant shows the ledges and parietal oscula in a normal con- dition on the dermal side. On the gastral side, which had be- come directly exposed to the outer world, the loose sponge- tissues had greatly increased, thus adding much to the thickness of the wall. The thickening had become in one part all the more considerable on account of the formation of a large cavity right in the middle of the resfenerated tissues. The cavitv evi- dently served the part of a gastral cavity newly formed. On the one side it is bounded by the old sponge- wall with its parietal oscula ; on the other, by a thinner wall consisting of the re- generated tissues, which are likewise perforated by a number of roundish gaps, the parietal oscula of the new formation. Some of these gaps are situated so close together as to form regular sieve-plate beams between them. The above case seemed note- worthy as illustrating the free formative plasticity dominant in the sponge-tissues. R. OKINOSEANA. — MISCELLANEOUS NOTES. 251 Of the inmate in the gastral cavity, I can only report that the single specimen (Mus. No. 487) with a complete wall, now before me, contains an Ophiuron ami a Polychœte Annelid, — no Crustacea. 252 I. IJIMA : HEXACTINELLIDA. I. REGADRELLA KOMEYAMAI nov. sp. PI. IX and PI. X, figs. 5-17. During 1898, Mr. Komeyama of Tokyo presented to the Science College Museum a very beautiful and remarkably well preserved specimen of what proved to be a new species of Rega- drella. I have named it in honor of the donor. The specimen was found by him at a collector's in Yokohama amongst other things that came from the Sagami Sea. I have no hesitation in assigning the locality of the specimen to that sea, though nothing whatever is known about the circumstances of its capture. The specimen (Mus. No. 486 ; PI. IX, fig. 1), which is in desiccated state, consists of two, — a large and a small individual, both thin-walled and lamp-chimney-like in shape. They stand close together being attached by means of large, irregularly lobed, basal expansions to a mass of soft, fine-grained tufa. More correctly, they have evidently not grown directly on the tufa, but on the basal mass of an individual of the same species long dead and destroyed. General Characteks. The larger individual, which is the better preserved of the two, will here be first described in detail. It measures 225 mm. in total length. . At the lower end the body is bent as if it had been directed upwards while growing with its base attached to a perpendicular surface. At the juncture with the knobby base, it measures not more than 30 mm. across. From that point superiorly K. KOMEYAMAI. — GEN. CHARACTEKS. 253 the diameter increases up to 75-87 mm. at about the middle. Thence upwards the body again gradually narrows to the region just behind the cuff, where the diameters measures 34-37 mm. The superior end may be called truncate, not rounded. Cross-section of the body presents an irregularly circular outline. The wall is 2-5 mm., at places only 1 mm., thick. The superior terminal osculum is 27-30 mm. in diameter. The entrance into it is guarded by a spicular wreath of corona as efficiently as by a sieve-plate. The corona (PI. IX, fig. 3) is composed of strong, straight or slightly curved, sharply pointed, spicular rays, which freely project out in a row from the angular oscular edge and stand out obliquely upward and inward to a length of 18 mm. or less. I count 39 coronal rays in all ; at the roots the intervals be- tween them are about 2'/2 mm. on the average. To the naked eye the rays present a peculiarly glistening appearance, which is due to the rough shagreen-like nature of their surface. Looking at the corona from above, its inwardly directed rays are seen arranged like the spokes of a wheel, leaving in the center but a narrow, free passage between their points. A coronal wreath of similar appearance has been known from Tœgeria pulchra described by F. E. Schulze in the Challenger Keport ('87). The same was later assumed by him ('95, p. 35) to have been mechanically produced by the accidental loss of the central part of a sieve-plate, such as is possessed by Dictyaulus elcgans. Whatever be its nature in T. pulchra, the corona in R. okinoseana is a perfectly natural feature. There can be no doubt about this not only from the presence of it in the second specimen, but also from the facts : 1) that the rays are invaria- 254 I. IJIMA : HEXACTINELLIDA. I. bly quite free of other spicules which would have remained sticking to them, had a sieve-plate been mechanically torn off; and 2) that their shagreen-like surface — a feature which is also shown by all the prostalia on the cuff-margin as well as on the lateral surface — has apparently arisen in relation to the free exposure of the parts thus characterized. At the same time there is no denying the fact that the corona had been ontogenetically derived from a sieve-plate. I assume that in an early developmental stage this comes into actual formation, if only to a partial extent, but that its component spicules are however soon loosened and lost as a normal process, leaving behind permanently only those that are deeply rooted in the lateral wall, viz., the coronal spicules. Genetically, therefore, the corona and the sieve-plate are to be considered as very nearly related structures, strikingly different though they appear to be. In this light the large terminal osculum should plainly and exactly correspond to the area which in certain other Hexac- tiuellids is covered by a sieve-plate. Since H. komeyamal and R. phœnix show a far-reaching similarity in other points of their organization, I am certainly not disposed to find in the corona of the former anything of more than specific value. The cuff is tolerably well developed in a continuous ring (PI. IX, fig. 3). Breadth up to 9 mm., as measured on the upper side. It is expanded outwardly and slightly inclined up- wards. While its superior surface is comparatively flat and well marked olf from the gastral surface by the angular oscular edge, the inferior surface slopes down to merge insensibly in the E. KOMEYAMAT.— GEN. CHARACTEKS. 255 general dermal surface. The ciifF is tlierefore thick at its base and becomes thinner tovvards its free, onter edge. From this edge there spring forth thin and inconspicuons prostal needles of varions lengths, forming a palisade-like, bnt irregnlar and mncli interrupted row. The longest of them may project to a length of 6 mm. They have, as already mentioned, rough surfaces. The parietal oscula (PI. IX, figs. 1, 2, 4) are round, 2-372 mm. ill diameter ; each bordered by a thin, iris-like membrane. They are tolerably uniformly distributed in right and left handed oblique rows. In the middle part of the sponge, they are situated at intervals of 6-18 mm. or more ; near the ends, more closely together. The external surface of the lateral wall may be said to be undulating on account of the low, flat and discontinuous swelling of the spaces between the parietal oscula. The swelling is by far too inconsiderable to be called a ledge, but culminates in irregu- larly conical or compressed elevations which are again small and never of any conspicuous height. Their summits, lying at intervals of 9-14 mm. from one another, bear each a group of thin, rough- surfaced prostalia lateralia before referred to. These j^i'oject to a maximum length of about 10 mm., and are arranged, in groups of only a very few or at the most of several together, in either closely adherent or diverging tufts. Otherwise, as when they spring from along the edge of a compressed prominence, they form a short row. The entire external surface is covered with a delicate, quadrate-meshed, dermal latticework formed of exceedingly fine beams (PI. IX, fig. 4). The sides of the meshes measure not more than 0.4 mm. in length. Each nodal point of the lattice- 256 I. TJIMA : IIEXACTINELLIDA. T. work is marked by a minute white spot, which under the micros- cope proves to be a fioricome borne on the tip of the distal ray of each dermalia. The meshes are seen to be overspread with a cribellate dermal membrane. Through the dermal latticework are distinctly visible the roundish or somewhat irregular-shaped apertures of the incur- vent canaU, which are of various sizes under 2 mm. diameter and are alwa3'^s rather shallow in conformity with the thinness of the sponge- wall. Between the said apertures the dermal lattice- work is in close contact with the parenchymal mass below. Amongst the latter, the trend of the coarser and more peripherally situated jiarenchymal bundles is traceable from the outside with sufficient distinctness. Arising from the compact base of the sponge, they run irregularly upwards in oppositely oblique directions, branching and uniting and loosely interweaving with one another without regularity. In places the bundles are fully 1 mm. thick ; more usually they are much thinner. The gastral surface (PI. IX, fig. 2) is devoid of a covering- latticework. The parietal oscula are seen on this side to occupy each a more or less depressed position, their iris-like membrane lying on a level with the general external surface of the wall. For the rest the gastral surface shows an uneven ness, firstly on account of numerous roundish excurrent apertures, and secondly, because of the most internally situated, coarse, parenchymal bundles which project in a ridge-like manner. Close to a parietal osculum, the excurrent canals are but ver}' small and shallow depressions. Farther away from it, they are much larger, often having a diameter of 2 or o mm. While some are pit-like thougli never ver}" deep, others are flat de- R. KOMEYAMAI. — GEN. CHARACTERS. 257 pressions into which two or more excnrrent canals open in common. The coarse parenchymal bundles, showing themselves on the gastral snrface, pursue an irregular, but on the w^hole transverse- ly directed course. Just inside the origin of the coronal rays along the superior oscular edge, there runs in a ring a strong, projecting bundle of fibers (see the upper part of fig. 2). Further, at this end of the wall a number of parenchymal bundles, running outside of the innermost, sinuously transverse bundles, assume a longitudinal disposition in association with the root of each coronal ray. If, in an imaginary case, all the finer and loose spicules should be removed so as to leave the coarser parenchymal bundles alone in situ, there would remain a frail, wide-raeshed and loosely inter- woven basket-w^ork, in which the onter and the inner bundles would be found, relatively speaking, to pursne directions inclined respectively in the longitudinal and the transverse directions. Thus, its general appearance would be much the same as in other species of the genus. However, one not unimportant peculiarity seems to consist in the fact that in the present species a much smaller portion of the skeleton at base, than] in either B. ohinoseana or R. phœnix, is subjected to synapticular amalga- mations. In the specimen described, the wall is quite firm in the region immediately adjoining the basal mass, which as usual is hard and compact. About 20 mm. above this region, the fusion of parenchymal spicules already becomes an occasional occurrence. A short distance farther above it is no longer to be found. I should think that, if all the loose spicules should be washed away, as they are after death on the native bottom, there would 258 I. IJIMA : HEXACTINELLIDA. I. remain with some degree of persistence only a basal cup of un- substantial structure, in height probably not more than one- tenth of that of the original specimen. With regard to the second, smaller individual, I may be brief. It stands out straight from the base without bending, the wall being more outbulged on one side than on the other. Total height 145 mm ; greatest breadth 63 mm. in one direction and 54 mm. in another. Breadth at inferior end 19-23 mm. Terminal osculum circular, 17 mm. in diameter. Cuff 2 mm. wide. Coronal rays not less than 33 in number, mostly 7-10 mm. long. Prominences on the dermal surface much less pronounced than in the larger individual. Spiculation. T\\Q ■parenchymalia consist predominantly of diactins, hexact- ins entering into their composition as occasional elements. The principalia in the parenchymal bundles are oxydiactins which may attain a length of 30 mm. and more \Yitli a breadth of 130// at the middle. Their form and arrangement agree well with the corresponding spicule in the specimen which will next be described, identified as R. phœn'ix ; but they are relatively more slender. The aceesoria, occurring as comitals or running either loosely or in strands by themselves, are mostly thin and filamentous diactins with an average thickness of about U //. and a length of 10 mm. or more. The center is cruciately knobbed, more com- monly only annulated. Subterminally usually sliglitly swollen E. KOMEYAMAI. SPICÜLAT10N. 259 and sparsely niicrotubercled ; extreme end rounded, sometimes conically pointed. Parenchymal hexactms of small or medium size occur only sparingly in the bundles of the wall proper. It required a study on sections in order to verify this fact. They occur more commonly in the cuff as well as in the parietal prominences (PI. X, fig. 11), in which parts they lie with one axis directed radially and in association with the proximal radial ray of the prostal hexactins soon to be described, into which they seem to merge by a gradational series of intermediate forms. — The paren- chymal hexactins are generally under 1 mm. in axial length. Not infrequently the radially directed axis in those situated in the cuff or in the parietal prominences is much longer than the others. Thickness of rays about 10 !>■ or less ; end generally bluntly pointed and sparsely prickled. The coronal spicule (PI. X, fig. 8) is evidently to be considered as a specially developed element of the parenchymalia. I may call it an oxypentactin with unequal rays, the sixth ray being represented by a mere boss. Its shape, position and manner of arrangement are essentially the same as in the corresponding spicule of R. phœnix (PL XÏ, figs, o, G). The same spicules in similar arrangement are also know^n from Dictyaulus elegans (Schulze '95). With respect to its longest complete axis the coronal oxy- pentactin is disposed longitudinally. That axis is more or less curved, the concavity facing inwards. The atrophied sixth ray is situated on the concave side. The superiorly directed ray of the longitudinal axis is the free coronal ray, the most strongly developed of all. In a large 260 I. IJIMA : HEXACTINELLIDA. I. specimen of the spicule that ray measured 18 mm. in length and fully half a millimeter in thickness at its base. It gradually tapers towards the finely pointed end. The surface is profusely equipped with strong, obliquely conical prickles (PL X, fig. 9) which cause the shagreen-like appearance when seen with the naked eye. Only small sections at the base and the end are smooth. As already noticed, the coronal ray springs out quite solitarily without the addition of other s]3icules, except occasional rhaphides which were found adhering to its surfoce. The four remaining rays of the coronal oxypentactin are all smooth. The inferiorly directed ray is always much shorter than its opposite, coronal ray. Accompanied with comital diactins it forms the longitudinal parenchymal strands visible for a short distance just inside the oscular edge. Still shorter than the inferior ray are the paired lateral and the unpaired outer rays. The former together with compactly set diactins forms the ring-like ridge just inside the origin of the coronal rays. The unpaired ray extends outwards into the cuft' which it transverses without pro- truding at the external edge. It thus affords a very efiicient support to the cuff. Being situated just under the superior culf surface, its course can be traced on that side as a whitish or a slightly raised streak proceeding radially from the origin of each coronal ray. The prodalia, both marginal and lateral, are the radially directed, distal rays of very variously sized oxyhexactins, which may be called the 'pvodal hexactlns (PI. X, fig. 11). These are, like the similarly situated stout dermalia in R. okinoseana (pp. 2oO, 235), linked to the parenchymal liexactins as well as to the ordinary dermalia by a gradational series of intermediate forms. E. KOMEYAMAI. — SPICULAÏION. 261 111 a large specimen of prostal hexactins the distal (prostal) ray, which is always the longest and the strongest of all the six rays, may be 10 mm. or more in length. The opposite proximal ray is shorter, and shorter still are the four paratangential rays. While the proximal and the paratangential rays are entirely smooth, the prostal ray is, like the coronal ray, beset with short, conical, obliquely outwardly directed prongs (PL X, fig. 10). In many cases, however, it was found smooth on one side — generally the concave side of gently curved shafts. In the smaller prostal hexactins, the above prongs are more weakly and sparsely developed. The hexactui-devmalia (PI. X, figs. 14, 15) are also subject to certain variations in regard to the size and proportional length of the rays. In general we may say that the rays are slender, measuring only about 9 /^- in thickness near the base. They are nearly entirely smooth or subterminally obsoletely rough ; the ends are rounded or bluntly pointed. The free distal ray varies in length from, say, 120/^ to 240/^. It is scarcely distinguishable from the other rays except by its relative shortness or by the fact that it is often slightly less tapering towards the end. The paratangential rays are straight or gently bent and about 300/^ in average length. The proximal ray may be either shorter (fig. 14) or much longer (fig. 15) than the paratangentials. The former is the rule es- pecially with those dermalia in which that ray ends free in the subdermal space, as is the case in such parts of the dermal layer as extend over incurrent apertures. The proximal ray is gene- rally longer — at times twice as long or even longer — than the paratangentials, and the entire spicule thus becomes more or less 262 I. IJIMA : HEXACTINELLIDA. I. sword-like, at places where the dermal layer touches the parenchy- mal mass, into which the ray in question enters like a nail. The dermalia are seen to extend over the membranous zone around the parietal osciila, forming the usual latticework right to the edge. Many of them in this situation have a very short proxim.al ray. In passing it may be mentioned that no more specially differentiated oscularia were found in this than in other species of the genus. A much larger size as given above is attained by some — not all — sword-like dermalia on the cuff-edge as well as on the summits of parietal prominences, in intermixture with the pros- talia occurring in these places (PL X, fig. 11). With the general growth in size of the entire spicule, the free distal ray especially undergoes elongation and development. It assumes a slender spindle-like shape, while the surface for nearly its entire length becomes roughened by the presence of microtubercles (Pi. X, fig. 18). I have measured such distal rays of 300 n, 400 /-«, 500 <« and more in length with a breadth of 22 ij. and over at the thickest part. Thus, as before said, the dermalia approaches, and finally becomes indistinguishable from, the smaller prostal hexactins. The gastralia, found in irregularly scattered distribution, are pentactins of moderate size, with the atrophied sixth ray indi- cated by a gentle swelling- The rays are straight or nearly straight, slightly tapering or uniformly thick (12 // or less) throughout. End almost always rounded ; subterminally obsoletely rough. The paratangential mys are often of unequal length in the same spicule; length up to 700/^. They run not always along the extreme gastral surface, being sometimes overlaid by R. KOMEYAMAI. — SPICULATION. 263 parenchymal diactins. The unpaired distal ray is usually much more elongated than the paratangentials. In exceptional cases, the gastralia seemed to be represented by thetactins or stauractins. TJie hexasters are floricomes, graphiocomes and onychasters, thus perfectly agreeing in this respect with R. phœnix O. Schm. The floricomes (PI. X, figs. 5-7) are borne in the well-known manner on the distal ray of almost every dermalia. Thej^ are also common in the subdermal region where they take origin. The diameter measures 136-152«. They are therefore consider- ably larger than in R. okinoseana or in the specimen which I identify as R. phœnix. The principal is 13 ," long as measured from the axial center and 3 thick ; it contains an axial canal extending to the very end which is slightly expanded. Five to eight terminals compose a perianth. This measures 13 ,« across the basal swollen part ; above the middle it narrows considerably (often to such an extent that the adjoining terminals almost touch one another), finally to expand to a width of about 50//. Each single terminal (fig. 6) is bosally very thin but as broad as 5// just behind the terminal disc. The latter, as seen from the top (fig. 7), is 11 mm. long and 15,« broad (teeth included). Its outer edge bears 5-S, moderately strong, recurved teeth. The rounded inner edge of the disc is plainly noticeable as such. Graphiocomes of the usual appearance are occasionally found, likewise in the subdermal region. Of much more common occur- rence are their detached terminals, the rhaphides, either scattered or still grouped in sheaves and found in various positions io the dermal layer (PI. X, fig. 12). An intact graphiocome measures 264 I. IJIMA : HEXACTINELLIDA. I. 200// in diameter. The principal is about 10// long, as measured from the axial center. The terminal sheaf is 12 /j: thick, keeping nearly the same width throughout or only sliglitly expanding toward the outer end. The onychasters (PI. X, fig. 12) occur in abundance every- where in the parenchyma. They are 80-9G /^ in diameter, i. e., about as large as in R. phœnix but of more uniform size. From each short principal (9 // long as measured from the axial center) there arise 3-5, thin, tapering, nearly straight and strongly diver- gent terminals. The finely attenuated end of these bears a whorl of 3 or 4, fine minute claws of just the same appearance as those to be described under R. phœnix (PI. X, figs. 20, 21). In some onychasters, the terminals and the claws were found to be less fine than in others. Finally, the rigid basal mass is composed of an irregular frame- work of siliceous beams (PI. X, fig. 17). These are 20-35/^ thick ; smooth but with occasional microtubercles or prickles. The inclosed meshes are, though not always, rather wide (measuring \\\) to 200// or more across). The framework is formed by synapticular fusion of not only diactin-parenchymalia but also of certain thick-rayed hexactins, the basidictyonalia, observed in so many other Lyssacina with hard bases (p. 232). This is proved by the presence of the hexradiate axial cross in the beams and also by such basidictyonalia in different stages of amalgamation as still retain their orio;inal external form. 265 REGADRELLA PHŒNIX O. Scum. PL X, figs. 18-27; and PI. XI. Regadrella phœnix, Schmidt '8o, p. Ol ; pl. VIII, figs. 6, 7.— Schulze '86, p. 39 (Rpr.).— Schulze '87, p. 84 ; pl. XIII, figs. 1-4.— Schulze '95, p. 34.— TopsEXT '96, p. 27ö ; pl. viif, fig. 1. — Schulze '99, p. 20 ; pl. III, figs. 3-0. Trichaiitella clegans, Filhol '85, p. 284 ; pl. VIII, Rhahdodictyum delicatum, Topsext '92, p. 25 ; pl. V, fig. 1. Begadrella phœnix O. Schm. lias been reported from tlie following localities : Near Lesser Antilles (Caribbean Sea) : Santa Cruz, 453 m.; Barbados, 404 m. & 52G m. (O. Schm. '80). Between Sta. Lncia and St. Vincent, 514 m. (F. E. S. '99). Near Azores : 861 m.) (Tops. '92), NW. of Sao Jorge, 1022 m. (Tops. '96). Bay of Biscay : 1410 m. & 1220 m. (Tops. '96). Coast of Morocco : 865 ni. (Filh. '85). Eastern Pacific : near Galapagos, 717 m. (F. E. S. '99). If I am not mistaken in referring to this species the speci- men, on which the following description is based, it is to be added to the above list : Coast of Chile, 3200 m. The said specimen is contained in the zoological collection kept for show purposes on board the U. S. Fish-Commission SS. ' Albatross.' During the two visits she paid to Japan in 266 r. TJiMA : hexactinellida. i. 1896 and 1900, opportunities for the study of that specimen were given me tlirough the courtesy of my friend, Dr. L. Stejneger, on the first occasion, and of Captain J. F. Moser on the second. The only information I coukl obtain about it was the statement on the label : ' A'^enus ' Basket, a siliceous sponge from 1800 fathoms. Station : Off coast of Chile.' A good description of an authentic and well-preserved R. 2^hœnix is still a desideratum. In fact only imperfect fragments of the species have as yet been studied with any care ; hence, the somewhat unsatisfactory state of our knowledge. In my at- tempts to identify the ' Albatross ' specimen, I found that it presented in its structure several points which seemed to be of importance as specific characters, but which were either uncer- tainly or not at all known from B. phœnix, or are perhaps quite wanting in that species. It is therefore with some degree of reserve that I refer the * Albatross ' specimen to B. phœnix. General Characters. The specimen (PI. XI, figs, 1, 2) is of a tubular form, torn off at the lower end. Length about 240 mm., representing, I should judge, nearly three- fourths of the original entire size. Diameter at the middle about 75 mm.; that at the lower end a])Out ~)0 mm. (It has previously been known that the species may attain a height of 500 mm.). The wall is thin, not more than .'3 mm. thick in the thickest part. It l)euds and falls in of itself Avhen taken out of the spirit in which it h preserved. The upper end is rounded, the lateral wall closing in slightly B. PHŒNIX. GEN. CHARACTERS. 267 all around toward tlie border of the flatly convex sieve-plate, which is encircled by an inconspicuous ridge-like cuff. I should remark that this state of the upper end seems to be usual in R. phœnix, at least after a certain stage in its growth. Schulze's figure in the Challenger Keport, taken from an au- thentic specimen given him by O. Schmidt, indicates just this condition of the upper end. And so do also the figures given by FiLHOL of his Trichaptella elegam, Avhich Topsent ('96) assumes to be identical with R. phœnix. Whether Topsent be correct or not in this assumption, it may nevertheless be pre- sumed that, as he had before him some perfect specimens of R. phœnix, he had observed an agreement in configuration be- tween these specimens and the above mentioned figures of Filhol. The sieve-plate in the * Albatross ' specimen is greatly damaged ; originally it must have been approximately circular and about 30 mm. in diameter (PL XI, fig. 1). Of its beams there remain only those Avhich must have formed the main sup- port of the plate, and which are themselves supported by the strong spicular rays that correspond to the coronal rays of R. homeyamai. The beams that remain project inwards from the cuffed border, are 2-6 mm. apart at the roots, and are arranged on the whole like the spokes in a wheel. Several of them meet or nearly meet at the center but without uniting in this position. However, there are some which in their inward course become confluent with their neighbors, thus forming triangular meshes. In a few places they show lateral branches, or rather remnants of these, which originally might have effected a continuous com- munication between two adjoining radial beams. — It scarcely needs to be pointed out that the beams are, unlike the coronal 268 I. IJIMA : HEXACTINELLIDA. I. rays in R. konieyamai, bundles of spicules, similar to the parenchy- mal strands. Their free ends bear unmistakable signs of having been forcibly broken off. Whether a central nodal plate, such as is possessed by Dictyaulus elegans (Schulze '95), originally existed but had been torn away, can not be ascertained ; but I think this much may be said, that the principal beams are in the main radially disposed and that the meshes are comparatively wide with angular corners. How far the radial arrangement of the main beams, consequent upon the strong coronal ray enter- ing into their support, can be verified in typical specimens of R. 'phœn'ix from the Atlantic remains to be seen. Parietal oscula (see PI. XI, fig. 2), circular in shape and mostly measuring 0-b mm. in diameter, perforate the wall at intervals of o-lO mm, arranged in two intersecting systems of irregular, oblique rows. They are thin-edged as usual and occupy each the center of a flatly depressed area bounded by the main strands of the parenchymal skeleton. Both the incurrcnt and the excurrent canals are visible on their respective sides of the wall as small and shallow depres- sions under 1 mm. in diameter. The dermal surface had been nuich abraded, exposing the more superficial parenchymal bundles. However, the occurrence of well- developed ledges or of hillocks with prominent prostalia must evidently be entirely denied. The surface is on the whole even, or more properly, gently undulating on account of the low and broad swellings between tlie parietal oscula. The dermal layer was found preserved in patches. In such places I have found the surface studded at intervals of 2-5 nun. K. PHŒNIX. — GEN. CHARACTERS. 269 with small irregularly papilla-lihe prominences, not more than 172 "1111. in height. These are evidently caused by the hydranth, invariably contained in them, of a commensal Hydrozoa which is harbored in the sponge- wall. The prominences were not observed in the uppermost region of the body only ; whether as the result of abrasion or not, could not be determined. Under the hand-lens, their summits as also the free edge of the cuff appear to be spiny (PL XI, fig. 3). The spines, protruding for not more than half a millimeter, are found to be the distal rays of certain specially developed dermal hexactins. Are the papilhe to be regarded as something of constant occurrence in the species ? In consideration of what we know about the relation of the Walieria species to the commensal Hydrozoa (see anon, under W. leucharti), this question is possibly to be answered in the affirmative. However, nothing like the papillîc, or the peculiarly modified dermalia (PL X, figs. 25-27) in connection with them, has before been described from R. phoenix. The coarse parenchymal bundles, exposed on the external surface, pursue a sinuous course in oblique or in nearly longi- tudinal directions (PL XT, fig. 2). The bundles, more deeply situated and exposed on the gastral surface, take a course which is inclined to be transverse in direction, similarly as we have seen in other species of the genus. In the upper part the principal bundles of the skeleton run obliquely right up to the sieve-plate border, exactly as is to be seen in the figure of E. phoenix given by Schulze in the Chal- lenger Keport. Some bundles, but by no means all, extracted from the lower 270 I. IJIMA : HEXACTINELLIDA. I. part of the specimen exhibit the synapticiihir fusion of the megascleric elements. Much less frequently such fusion is to be met with in about the middle, and none at all still higher above. Were the base of the specimen preserved, just the same condition as has been known from R. phœnix in regard to the fusion of spicules in that region would undoubtedly have been found. The said condition of the base admits, as was remarked by Schulze ('87), of the occurrence of several specimens growing one within another, each inner one representing a younger generation seated within the dead skeletal remnant of the pre- ceeding generation. O. Schmidt ('80) observed such occurrences when he first described R. phwiiix, which specific name he chose on that account. Topsent ('96) mentions a case in which as many as fine generations were represented in the manner noticed, and recently Schulze ('99) has added still anotlier instance to the list. It scarcely needs to lie remarked that this piling up, as it were, of successive generations, forms no specific peculiarity of the species. In R. okiiioseana I have found it a very com- mon occurrence that the young specimens are attached to the dead skeletal stump of the same species. So also the two indivi- duals I have described of R. homeyamai are attached, not one within the other, but close together, side by side, on the basal mass of an individual long dead and destroyed. SnCLLATION. The spiculation shows an especially close agreement with that of the species last describ 'd. Excepting the large oxypentactins common to the sieve-plate R. PHŒNIX. SPICULATION. 271 and tlie lateral wall, the parenchymalia are chiefly cliactins with an occasional sprinkling of slender thetactins and of similar hexactins which have one axis greatly elongated in excess over the others. While the thinner parenchymal strands consist solely of the thin filamentons spicules, the coarser ones contain large bow- shaped or boomerang-like oxydiactins, the principalia, in addition to the much more slender comitalia. The principalia may measure 50 mm. or more in length and 290/^ in thickness at the middle which is smooth and either gently curved or bent in an elbow-like manner. The finely attenuated ends are smooth-sur- faced. In forming a bundle, the principalia are arranged side by side and one after another in overlapping series, each sur- rounded by a copious quantity of the comitalia. The comitalia and all other fine parenchymal diactius are either annulated or cruciately knobbed at the center ; their ends rounded or conically pointed, often swollen and subterminally roughened to a greater or less degree. The radial beams of the sieve-plate, which are to be con- sidered as outward continuations of the parenchymal bundles, are supported by certain prolonged rays of large, unequally rayed oxy- pentactins (occasionally oxystauractins) arranged in a circle along the sieve-plate border. (PI. XT, figs. 5, 6). Similar spicules in the same position and arrangement have been described by F. E. Schulze in Dictyaulus elegans, and by myself in R. home- yamai (see p. 259). The plane of the two corap)lete and cruciately disposed axes is concave on the inner side, and on this side the sixth atrophied ray is always represented by a small, conical protuberance. The outwardly directed, unpaired ray is 272 T. IJTMA : HEXACTINELLTDA. T. relatively very short (not over 2 mm. in length) and is some- times represented merely by a rounded boss. It enters into the support of the cuft*. Of the cruciate, paratangential axes, the one in longitudinal direction is the longest, reachiug up to 30 mm. in length. The superior ray in this axis, i. e., the ray which goes into the com- position of each of the radial sieve-plate beams, is in length about equal to or sliorter than the inferior ray which is imbedded in the parenchyma of the lateral wall. The lateral rays in the transverse axis are always much shorter than either the superior or the inferior ray ; sometimes the}'^ are as short as the distal unpaired ray. Thickness of the rays at base 250-475 ,«. The superior and the inferior rays commonly thicken somewhat at a short distance from the central node, then gradually narrow again toward the finely pointed end. The lateral rays, lying along the base of the cuff, run in association with just the same diactin elements as compose the parenchymal bundles of the lateral wall. The superior ray is distinguished from all the rest by having a number of obsolete microtubercles widely and sparing- ly distributed over its surface (PI. XI, fig. 7). The microtubercles are frequently only indicated. They disappear entirely to- wards either end of the ray. Along with the ray in question are found in a bundle bow-like oxydiactin-principalia and diactin-comitalia, to complete the parenchymalia of the sieve-plate (PI. XI, fig. 8). Among the comitalia are not uncommonly found small and slender-rayed hexactins, which jDass over into the shorter diactin-comitalia bv a 2;radational series of inter- mediate forms. In the Challenger Report F. E. Schulze gave large oxy- R. PHŒNIX. — SPICULATION. 273 pentactins as tlie principal parenchymalia of the species. Later ('99) he described for these strong oxydiactiiis bent in a boomerang-like manner. This apparent contradiction may be explained by assuming that on the first occasion he had before him the above-described oxypentactin of the marginal zone, which assumption is all the more likely since at the time only the upper end of the body was available for his study. The dermalia (PL X, figs. 23, 24) — as they occur on the general surface, forming a delicate, quadrate-meshed latticework, the meshes measuring only about 240/-« in length of sides — are small, slender-rayed hexactins in which tlie distal ray is the shortest and the proximal ray, usually but not always the longest of all the rays. The former is only GO-90 /.« long and scarcely ever exceeds G// in thickness. It slightly broadens towards the rounded outer end, but never to such a marked degree as to give it a distinctly club-like form. The surface is rather sparsely provided with almost obsolete microtubercles. All the other rays taper gradually toward the bluntly pointed end. Tiiey are smooth all over or subterminally faintly rough ; straight or some- Avhat bent. The paratangentia.ls are 150-230 /i long. The prox- imal ray is, as indicated above, usually much longer (fig. 23), but sometimes only about as long ; it may even be considerably shorter (fig. 24), as, for instance, in those dermalia situated on the membranous edge of the parietal oscula. Dermalia of the above description have been hitherto un- known from the species and constitute one of the points which it is exceedingly desirable should be tested in specimens from the Atlantic. 274 I. IJIMA : HEXACTINELLIDA. I. The dermalia present tliemselves in sjiecially large size and strong development on the papillœ which as we have seen stand in relation wdth the hydranth of the commensal Hydrozoa, as well as on the free edge of the cuff (PL X, figs. 25-27 ; PL XI, fig. 4). They give to these parts the spiny appearance already mentioned and might not improperly be called prostal hexactins, were it not for the comparatively insignificant length attained by the freely projecting, distal ray. On the papillœ (PL XI, fig. 3), the dermalia are sword- shaj^ed hexactins, many times larger than the ordinary dermalia and with the distal ray swollen into a fusiform or a club-like shape (PL X, figs. 18, 25-27). This ray generally measures 600-870/^ in length and 45-80 /^- in greatest breadth. Its lateral contours are frequently not symmetrically even. Close to the rounded or bluntly pointed end the surface is rough on account of the presence of either pointed or rounded microtubercles. Sometimes the ray is smooth nearly all over. All tlie other rays (11-27 /-« thick near base) are tapering, subterminally spar- ingly microtubercled, and end couically or obtusely pointed. The paratangentials are comparatively very short and often of unequal length in the same spicule. The long proximal ray, which dips into the parenchyma accompanied with diactin- comitalia, may be 2 mm. or more in length ; occasionally as short as the paratangentials. It is not infrequently more or less distinctly bent in its course. From PL X, fig. 22, or by comparing figs. 23-24 with figs. 25-27 in the same plate, will be obtained a fair idea of the difference in size between the ordinary dermalia and those in a group around the Hydrozoan body. The paratangentials of the latter kind of dermalia generally lie somewhat below the general R. PHŒNIX. — SPICULATION. 275 level of the dermal layer. I tliink the two kmds are connected by forms of intermediate shapes and sizes. Of specially strong development are also the dermalia which lie crowded on the cuff-edge. Here they are again sword-like hexactins (PI. XI, fig. 4), differing from the papillär dermalia in being somewhat stronger on the average and in having the distal ray slightly differently characterized. They may measure in total length 3 mm. or more, of which 1 — 17^ mm. belong to the distal ray. This ray may reach 100 <« in thickness at base. It is tapering, though the lateral contours are not always even or straight. It bears low and somewhat scaly tubercles which are either confined to the end or extend over its outer half. In the latter case they are notably smaller and more crowded to- wards the bluntly pointed end of the ray. Paratangential rays may be as long as 800 /^ ; their pointed ends rough or nearly smooth. Both Schulze and Filhol have figured bristle-like prostalia projecting to a length of several millimeters on or around the cuff. I have not found the like in the specimen examined by me. The said prostalia are, according to Schulze, the free distal ra^^s of hexactins ('87, pi. XIII, fig. 2), which resemble some of the marginal dermalia I have seen, except that they are much larger. The gadralia are pentactins somewhat larger than the ordi- nary dermalia and distributed without regularity in their arrange- ment. The paratangential rays are usually unequally long in the some spicule ; up to half a millimeter or more in length ; about 10 IX thick for the greater part of their length ; more or less curved ; subtermiually somewhat swollen and rough ; end rounded. 276 I. IJIMA : HEXACÏINELLIDA. I. The unpaired ray is usually much longer and more pointed at the end than the paratangentials. Of the hexaslers, the floricome (PI. X, figs. 18, 19) is found in abundance in the periphery of the wall, now and then borne on the end of dermal distal rays in the usual way. Diameter 100-115//, which size nearly corresponds to that of the same rosette taken from O. Schmidt's type specimen and figured by Schulze in the Challenger Keport. The disc at the end of each principal is convex on the outer side. Terminals 5-8 in a perianth. Terminal disc with 4 or 5, moderately strong teeth on the outer edge ; the inner edge in profile view is indicated by a hump-like bend of the contour-line at that place. The graphiocome in an intact state was rarely observed, but the rhaphides detached from it were found in tolerable abun- dance, either scattered singly or still preserving their sheaf-like arrangement. They mostly adhere to the dermalia ; otherwise, they lie about free in the most peripheral region of the wall. Length of rhaphides 95 lu Kelics of the graphiocome, consisting only of its principals with discs at the ends, have been occasion- ally met with. Principals 10 /-« long ; rather slender. Except in a specimen from the Galapagos (F. E. Sch. '99), the graphio- come seems to have hitherto been overlooked. The onychasier (PI. X, figs. 20, 21) closely resembles the same of R. Jcomeyamai. It is very abundant, especially in the deeper parts of the wall. Diameter 64-92 /i. The fine, tapering terminals number o or 4, rarely 2 or 5, to each principal. The short, exceedingly line, backwardly arched, terminal claws seem to number o to each terminal ray. They may be easily over- K. PHŒNIX. — SPICULATION. 277 looked, and the rosette taken for an oxyhexaster, unless a Ligli power be used in examination. Nobody lias given the measurement of the onychaster from Atlantic specimens. However, by computation from the scale of the figures given by Schulze and Topsent, the diameter should be 64-90 /i, a range of variation well agreeing with that in the * Albatross ' specimen. 278 I. IJIMA : HEXACTINELLIDA. I. Walteria F. E. Sch. So far as known at present two species make up this genus, viz., W. flemmingi F. E. Sch. and W. leucharti Ij. The generic diagnosis would be : Euplectellids of saccular or tubular shape, firmly attached by the expanded base; with oscula on the sides; th^ surface with simple or branched processes (produced by a commensal Hydroid colony). Principal parenchy- malia, diactins; intermedial parenchymalia of small, spinous oxyhexactins. The hexasters are: Floricomes (which may be wanting) ; spherical discohexasters Avith numerous terminals ending in an arched disc; onychas- ters; and graphiocomes. The spiculation indicates the close affinity of the genus to both Tœgeria and Regadrella. The differential characters of the two species are : a. Saccular, the wall consisting of a wide-meshed latticework of beams. With floricome. Spherical discohexaster about 62 fx in diameter... W. fleinmlngl (N. of Kermadec Is.). b. Tubular, with numerous side-branches giving a tree-like form. Without floricome. Spherical discohexaster 75-00,« in diameter W. leucharti (Sagami Sea). 279 WALTERIA LEUCKARTI Ij. Pis. XII & XIII. Walte^'ia leuckarti, Ijima '96 p. 251. Hyalodendron navalium, Moore '98. The exquisitely tree-like Euplectellid from Japan recently described and figured by J. P. Moore under the designation of Hyalodendron navalium n. gen., n. sp., is indubitably identical with my Walteria leucharti previously described in the * Zoolo- gischer Anzeiger.' The brevity of my description may have had much to do in causing Moore's failure to recognize the identity. The creation of a new genus to receive the species is, in my opinion, inadmissible. The description given by Moore is on the whole good, but requires some important additions and correc- tions. Numerous specimens have passed through my hands. The exact localities in the Sagami Sea, where specimens of the species have been obtained by Kuma, are as follows : Outside Okinose, 717 m. (500 /m'o = 392 fms.). Near Mochiyaraa, 1000 m. (700 /wVo=ô46 fms.). Homba, 500-572 m. and over (274-313 fms. +). Gokeba, 572 m. (400 /wVo = 313 fms.). Fragments and grains of tufa attached to the basal disc attest the nature of the bottom. Once at Gokeba, a haul of the long-lines brought up large fragments of four different individuals at a time. From Outside 280 I. IJIMA : HEXACTINELLTDA. T. Okinose I have a specimen consisting of two individnals whose stems had come in contact crosswise and had fused tooether. These cases may indicate that the species grow close together side by side in certain localities. The species has as yet never been obtained in the Sagami Bay, north of the Okinose ridoe. General Characters. The three specimens shown in PI. XII, reduced to 74 natural size, will give a good idea of the general appearance of the sponge. It resembles in a measure a Cryptomeria or a fir-tree denuded of leaves. The body may be said to consist of the basal disc, the stem and the lateral branches. The ha&al disc is large, solid and compact. It may measure 120 mm. or more in diameter and about 3 mm. in thickness at the blunt-edged margin. The thickness increases towards the origin of the stem. The disc may be irregularly shaped con- forming itself to the character of the rocky substratum. On the superior surface there are usually seen in small numbers and in indefinite positions thin and sharp edged openings, which may be as large as the oscula on the stem but usually are smaller. They lead into shallow cavities, the wall of which may again show perforations penetrating for some distance into the hard basal mass. These are evidently excurrent canals, the external openings being undoubtedly oscula. This leads us to assume that the flagellated chambers occur even in the disc and that the circulation of water takes place here in a manner similar, W. LEUCKAETT. — GEN. CHAKACTERS. 281 to that in tlie upper part of the sponge, though probably with less energy. From near the center of the basal disc arises the stem, the sponge-body proper, which is tubular. This is usually more or less bent. In large specimens it may reach a height of one meter. Its thickness at its base may about equal that of one's finger, but is somewhat thinner in average specimens (11-12 mm. dia.). As a rule it slightly narrows above for a short distance, thence either to keep up a nearly uniform caliber or to apprecia- bly thicken again (up to 17 mm.) towards the middle section of the stem, where the lateral branches are in strongest develop- ment. Further above, the stem shows a gradual tapering to the apex. I find this end broken off in most cases. In the speci- men of PI. XII, fig. 3, it is preserved, showing that the apex tapers off to a point and is closed. The cross-section of the stem is on the whole circular in outline ; sometimes rather polygonal. A well developed lumen, i. e., the gastral cavity, extends through the stem from base to apex. Thickness of wall in the middle of body 1.5-2.5 mm. Towards the base the wall becomes much thicker at the expense of the caliber of the internal lumen. The moderately large oscula, to be seen here and there on the wall, are sometimes round but more generally oval or elongate oval apertures with the longer diameter disposed in the longi- tudinal direction. They are by no means uniform in size ; the largest may measure 10 mm. in the longer diameter. The oscular margin is sharp-edged, scarcely thicker than a sheet of paper and is usually raised into a low crateriform or lip-like rim. The distribution of oscula on the stem is quite irregular. 282 I. IJTMA : HEXACTINELLIDA. I. Sometimes but a narrow bridge-like space intervenes between two adjacent oscula ; more frequently are they situated more than 100 mm. and at times even 200 mm. apart on the same side of the stem. Their total number is therefore never very great. In a large specimen with a height of 855 mm., I have counted iu all not more than 25 oscula on the stem. The u^Dpermost osculum may lie within a distance of 10 mm. or so from the apical end. Inferiorly, an osculum may occur right at the stem-base and, as before mentioned, even on the basal disc. There apparently exists no definite relation between the distribution or size of oscula and the develo2:»ment of lateral branches. Seen with the naked eye or under a lens, the surface of the lower stem-end shows the same structural character as the basal disc, — that is to say, a densely interwoven texture of fine spicules intersecting in all directions. On these parts, a loose superficial tissue is usually entirely wanting. Such a tissue generally begins to exist on the stem a few centimeters from the base and covers the rest of the sponge parts in a thin, but by no means uni- formly distributed, layer. This tissue, together with warty protuberances on the stem and the branches, gives to the sponge an appearance fittingly described by Mooee to be, ' as if the specimen had been dipped into a thick soap lather, which had been allowed to dry on its surface.' Through that covering layer can be plainly observed the outer spicular bundles of the parenchyma, traversing the stem on the whole in a longitudinal direction. They are of varying thickness and closely set, frequentl}^ uniting and again separating in their course or sometimes intersecting one another at low ansfles. W. LEUCKARTI. — GEN. CHABACTEES. 283 On the other hand, the gastral surface of the wall (see PI. XIII, fig. 5) shows the inner system of weaker and more widely set parenchymal strands which are directed in the main trans- versely or obliquely. Thus it will be noticed that the general arrangement of the parenchymal strands is the same as I have described for Regadrella. The small irregular meshes on the gastral surface, formed by the intersecting of parenchymal strands, are each occupied by a shallow or pit-like depression, the excurrent canal. What now give the most characteristic feature to the species are the branches and wart-like tuberosities, which latter occur on both the stem and the branches but more numerously on the branches. Let it at once be stated that the branches arise by growth from the wart-like tubercles, and that the production of both is evidently dependent upon the commensal Hydroid colony tenant- ing the sponge-wall, as has been maintained by F. E, Schulze also in the case of certain tulnilar structures in W. Jlcmmingl. The branches in W. leuckarU are in a sense comparable to, although genetically different from, the ledges of J^ujjlectella, the stem being the most essential part of the body. The tubercles are sometimes low and not at all pronounced ; sometimes distinctly wart-like or even tubular. Seen under a lens, they present a hispid exterior on account of the dermal hilt-rays. (PI. XIII, figs. 20, 21). Their general aj^pearance, especially as they stud the branches, reminds one of the polyparies of a Madreporarian skeleton, and that all the more, since each tubercle has a small opening in its summit. The more pro- minently developed tubercles have more than one opening besides 284 I. IJIMA : HEXACriNELLIDA. I. the terminal one. Each such opening leads into a small cavity (PI. XIII, fig. 21 ; cav. hy.) which invariably harbors a hydranth of the commensal Hydrozoa. The cavity serves the part of a hydrotheca to the naked hydranth. Special canals proceeding from it for the reception of the branched cœnosarc can not be perceived. According to Kuma's statement, the tips of the tubercles in the fresh state present a pinkish color, which probably belongs to the Hydrozoa in question. There exist on both the stem and the branches all sorts of transitional stages between simple tubercles and such as deserve to be called inceptional branches or twigs. The growth into the latter is evidently initiated by a multiplication of the hydranths and by the formation of new openings for these, and eventually of new tubercles. So long as they are small and simple, the tubercles are made up entirely of the loose, superficial sponge- tissue and can therefore be easily rubbed off. As they progress in their development into branches, they acquire for their support an axial core of parenchyQial strands branching out from the same of the parts on which the developing branch is borne (see the lower end of fig. 20). PI. XII, fig. 2 shows a skeleton in which the superficial loose tissue together with the tuberosities had been entirely scraped off. The branches, varying in thickness from less than 1 mm. to oY^. mm., spring out on all sides of the stem, though exhibiting different degrees of development in certain pai'ts as will soon be pointed out. They generally arise from the stem at nearly right angles. When somewhat obliquely inclined, they are directed superiorly about as often as inferiorly. They are usually nearly straight, though I have found many of the branches almost uni- formly curved upwards in onu specimen and downwards in another. W. LEUCKARTI. — GEN. CHAKACTEES. 285 The branches enianatmg from the stem may be simple and unbranched, in which state they may sometimes attain a length of 40 mm ; but more usually are they provided with secondary branches, developed in number and length proportionally to the primary branch, which, when large, may bear even tertiary branchlets on its secondaries. The secondaries and tertiaries shoot out more or less inclined towards the apex of the branches bearing them, frequently forming with the latter an angle of about 45°. On the lower portion of the stem, the branches are obliterated and are represented by compact stumps. The longer stumps may be frayed out at the outer end into a tuft of separate needles. Such remnants of the branches are sometimes found even on the basal disc. Complete branches begin to exist at some distance from the basal end. They are at first all small. Larger and more complex branches add themselves to the smaller, in a generally gradual development, towards the middle of the stem. Thence tovvards the upper end the branches again become continually shorter, and finally they become so very short that the general form of the sponge gradually narrows towards, and is pointed at, the apex (PI. XII, fig. 3). In a very large specimen (855 mm. high), one of the largest branches was 200 mm. long, bearing numerous secondary branches up to 55 mm. in length. I have said the branches arise on all sides of the stem ; but it must not be supposed that they are equally developed in all directions. As a constant feature seems to be the situation of the larger branches oppositely along two sides of the stem. As the result of this arrangement the entire sponge is of a more or less flattened form : it is laterally compressed, if one may so express it. It is also distinctly noticeable that the secondary branches 286 I. IJIMA : HEXACTINELLIDA. I. are best developed in the plane of the primary branches bearing them. The symmetry and regularity in form of the entire sponge is however usually much disturbed by the bending and twisting of the stem as well as by the not uniform development of the lateral branches. In certain specimens, as an occasional occurrence a branch may be swollen, either terminally or elsewhere, into an irregular mass of considerable thickness. For instance, the specimen of PL XII, fig. 1 shows such swellings in at least two places. Cutting one open, I found no free cavity within but a space traversed by loose parenchymalia, in which was imbedded the body of a small animal — probably an Annelid. I think the swellings are always mere abnormalities caused by certain extrinsic objects. The process of their formation may in a measure be compared to that by which all the branches arise in connection with the commensal Hydrozoa and yet the latter comes near to being intrinsic on account of its invariable presence. In some specimens it is not at all uncommon to observe a thin web-like expansion of the spicular tissue at the axils of the branches. One specimen which I have seen was jDarticularly distinguished by the great abundance of the web-like plates not only on the branches themselves but also between these and the stem. On the other hand, a number of specimens nowhere showed a similar development. Spicülation. The jKirenchymaUa are almost exclusively diactins of vari- ous sizes. Very rarely among the accessor la there occur spicules W. LEUCKAETI. — SPICULATION. 287 with a greater number of rays, as, for instance, hexactins witli one of the axes elongated considerably more than the others. The oxydiactin prmcipalia may measure 35 mm. in length and 120/^ in breadth at the middle. They are usually nearly straight and smooth throughout. They are present as usual in all sizes, grading down to fine accessoria which occur either as comitals or singly by themselves. Tlie accessoria are of only 7-11 fji breadth, either plain or annulated at the center, and sub- terminally swollen and rough-surfaced, the extreme end being rounded or obtusely pointed. At the juncture of a primary branch with the stem, the axial parenchymal bundle of the former joins the outer bundles of the latter. At the root of the branch some of the fibers pene- trate for only a short distance into the said parietal bundles ; others are seen to spread out in all directions among these. A similar arrangement obtains at the origin of secondary and tertiary branches. The coalescence of the parenchymalia by simple fusion as well as by numerous synapticula? is carried on to a great extent in the stem. Only for a short stretch at the apical end the paren- chymalia are all loose. The ankylosis is especially dense nearer the gastral surface and towards the basal end. It extends into the base of the primär}^ branches and often farther outwards, but seldom into the secondary branches. The greater part of the basal disc is composed of parenchymal diactins disposed not in strands but rather in a feltwork-like arrangement. They are compactly soldered together in a close- meshed framework. The compactness increases from the superior surface inwards. On the inferior surface, in direct contact with 288 r. iJiMA : hexactinellida. i. the substratum, there is the usual reticular plate. This is thin and has small roundish meshes, measuring 20-40 /^ across. Scattered among the parenchymal megascleres are found small, slender-rayed and spinous oxyhexactins (sometimes oxy- pentactins by suppression of a ray). (PL XIII, figs. 18, 19). Doubtless we have here the same intermedial microxyhexactin which is known from Regadrella ohinoseana (Pi. VIII, figs. 24- 20), Tœ(jeria indchra and Dictyauhis elegans. The spicule in question generally measures 190-260 ^^^ in axial length, the rays being only about 4 n thick close to the base. Numerous small spines beset the entire length of the rays ; they are directed somewhat obliquely outwards, though there exist others which stand out nearly or quite vertically. They become obsolete to- wards the finely pointed ends of the rays. The oxyhexactins thus characterized are not very numerous in the stem or in the branches. They occur in greatest abundance in the upper super- ficial layer of the basal disc. The (Jermalia (PI. XIII, figs. G-9, 21) are exquisitely sword- like hexactins wdiich are closely and rather indiscriminately set together, so that the paratangentials of separate dermalia do not form a regularly meshed latticework nor are they arranged all in the same level. They vary considerably in respect of size and of the relative development of their several rays. Those of the larger size measure 1 mm. or somewhat more in total length, the proximal blade-ray being five or six times as long as the distal hilt-ray. Such a large size is attained especially by those der- malia which enter into the composition of the wall of the wart- like tubercles. This reminds us of the tubercles of similar W. LEUCKARTI. — SPICULATION. 289 nature and structure in Regadrella phœnix (p. 274). In other places the dermalia are usually much shorter— sometimes only ahout half as long, the blade-ray being of about twice the length of the hilt-ray. Thickness of rays at base 4 /^ in the slender-rayed forms, but as much as 10 /^ in the more stoutly developed ones. The hilt-ray, 165-250/^ long, gradually thickens distally, finally to contract again and to terminate in an acute or obtuse point, rarely in a rounded knob. Thus, it is of a slender club- like shape. In a large specimen of the spicule, the swollen part may be 20/^ thick, i. e., about twice as broad as at the base. The surfiice is rough on account of low scaly microtubercles which gradually lose themselves toward the base of the ray. The roughness may be obsolete, especially on the more slenderly developed hilt- ray. The paratangential rays, 80-200 ix long, are usually slightly broadened toward the end, which is sparingly microtubercled and obtusely pointed. They are not always quite straight, nor of the same length in the same spicule. The proximal blade-ray tapers towards the pointed and faintly roughened end. It is not infrequently bent in adaptation to the circumstances of its occurrence. As gastralia (PI. XIII, fig. 10) are to be considered pentac- tins of medium or small size, found isolated and by no means numerously on the inner surface of the stem-wall. The paratan- gentials are usually 90-145 F- long and about 8 /^ thick at base, while the distally directed, unpaired ray is about twice as long or longer. All the rays are smooth except near the conically or bluntly pointed end. A knob on the proximal side of the spicular center represents the atrophied ray. Moore's statement 290 I. TJIMA : HEXACTINELLIDA. I. ('g8, p. 433) that the gastralia, are sword-shaped hexactins like the dermalia is not corroborated by the facts. Tlie hexasters of the species are the spherical discohexaster, the onychaster and the graphiocome. Remarkable is the absence of floricomes, which should be present in W. flemmingi in addition to all the three kinds of hexasters above-mentioned. The spherical discohexaster (PI. XIII, figs, 12-14), which is of great beauty and closely resembles the same of W. flemmingi in appearance, is found in great abundance, especially in the periphery inside the dermal paratangentials. Frequently it is found also outside these and sometimes even borne on the tip of dermal hilt-rays after the manner of floricomes in other Euplec- tellids. In deej) parts of the sponge the discohexaster is wanting, or at any rate quite scarce. In diameter the discohexaster measures 75-90 /^. Each short and rather slender principal bears at the outer end a plano- convex or nearly hemispherical disc (fig. 13). From all over the outer arched surface of this disc spring out numerous termi- nals, which are slender at base but gradually thicken outwards, finally to end witli a watch-glass-like or hemispherical, marginally minutely serrated, terminal disc of about 4/^ diameter (fig. 14). It is difiicult to determine the number of the terminals. In well developed cases, there must be to each principal thirty or more of them in a diverging bunch. In each bunch the more peri- pherally situated terminals are gently curved at base, while the central ones are straight. All the terminals diverge outwards in such a way that the terminal discs are uniformly distributed over the entire surface of the exquisitely spherical hexaster. W. LEUCKARTI. — SPICÜLATION. 291 The delicate ony chaster (PI. XIII, figs. 15-17) is of inucli the same appearance as in W. flemmingi. It is rather rare, at any rate not common in all parts. While in most preparations it required a prolonged search to discover one, in others several were found side by side among the parench3^mal diactins. This gave me the impression that the onychasters were in the main more deeply situated than the spherical discohexasters, though sometimes both occurred in intermixture. Diameter 68-84 ji. The principals are very short, thick and swollen at the ends ; each bearing 3-6 (rarely more), exceedingly fine and strongly divergent terminals. These arise from the principals without regularity in arrangement, not in a whorl (fig. 16). They generally taper towards the outer end, which is capped by a whorl of gently recurved, minute claws, usually 4 or 5 in number. The cap, though somewhat more strongly developed in some cases than in others, is ordinarily so small as to require careful observation under a high power in order not to overlook it. In several instances the terminals revealed no claws even when examined under the immersion system, but seemed actually to end with a minute pinhead-like knob. Moreover, the excessively fine outer ends of the terminals easily break off; so that the chances of the hexaster in question being erroneously taken for an oxyhexaster are great. Moore seems to have fallen into this error. The graphÂocome (PL XIII, fig. 11) is of typical form. It is very large, reaching almost 450 /^ in diameter. Sheaf of rhaphidial terminals 200/^ long and 20// broad; the central, hexradiate principals 26 /^ in axial length. As usual the graphio- comes are confined to the periphery of the sponge. In the perfect 292 I. IJIMA : HEXACTINELLIDA. I. state they are but seldom seen. Common however and in places quite abundant are the more or less disintegrated terminal sheaves which have separated from the mother-rosette and taken up a superficial position, lying vertical to the external surface along with the radial rays of the dermalia. Moore has failed to recognize the graphiocome in its entirety. By him the terminal rhaphides have been mentioned, as ' acicular diacts,' while its central part which remains after the loss of the terminals has been taken by itself for a special kind of rosette. Miscellaneous Notes. All the specimens at my disposal were not in a fit state for a study of the soft parts. It could however be determined that the chamber-layer extends to the extreme end of the branches ; further, that the ectosomal surface, lying some distance above the dermal paratangentials, is lifted into little conuli by the ends of the dermal hilt-rays (PI. XIII, fig. 21). PI. XIII, fig. 4 shows a case of monstrosity in which the superior end of the stem is dilated into an irregularly pyramidal, compressed and thin-walled sac. The wall is perforated by a number of typical oscula. This anomaly had evidently arisen in connection with the rej)arative growth after the stem had sus- tained an injury in that part. AVith respect to the commensal — possibly symbiotic — Hydro- zoa, the state of preservation was in no case such as allowed an exact investigation into its characters. But this much could be ob- served : that each hydranth possesses nujuerous finger-like tentacles W. LEUCXAETI. — MISCELLANEOUS NOTE. 293 and inferiorly passes, without sharp demarcation, into a stalk- like portion of the branched cœnosarc, which traverses the parenchymal tissue of the stem-wall and of the branches. The cœnosarc is 30-60/^ thick and apparently solid, though probably in fact tubular. Its branches possibly undergo anastomoses. In places it seemed to bear shorter or longer, blindly terminating branches which were sometimes swollen and club-like at the free end. These are probably early stages in the development of new persons by budding. A perisarc is wanting to the entire colony (gymnoblastic). Nestle-capsules oval-shaped and very small, measuring scarcely 5 ;^ in length. The Hydrozoa is at any rate closely similar to the species inhabiting W.flemmingl, which has been figured by F. E. Schulze in the Challenger Keport, PI. XI, fig. 4. It is clearly different from either Stephanoscyphus mirabilis Allm {=Spongicola fistulosa F. E. ScH., known to inhabit several Monaxonid species) or Amphibrachiwn euplectellœ F. E. Sch. ('8o). F. E. Schulze, when describing the external from of W. ftemmingi, had some doubts as to whether he had to do with a normally shaped specimen, or not rather with one essentially modified on account of the presence of the commensal Hydrozoa. He was apparently led to entertain this doubt from a certain Adriatic Myxilla which is normally of a compact bulbous body, but acquires a shape like a tuft of the common heath when invaded by Stephanoscyphus mirabilis. As for W. leucharti, I think the shape ascribed to it in the above, and its association with the Hydrozoa, may fairly be said to be constant, since, in more than a score of specimens, not a single case has been observed that suggested the contrary. 294 I. IJIMA : HEXACTINELLIDA. I. Literature referred to in this Contribution. Max Schultze. '60. bowerbank j. s. '62. Schmidt, 0. '64. Barboza du Bocage J. V. '65. Herklots, J. A., & Marshall, W. '68. Thomson, C. W. '68. Wright, E. P. '68. Barboza du Bocage, J.V. '70. Kent, W. >S. 70. Schmidt, 0. 70. Thomson, C. W. 70. IIaeckel, E. 72. Bowerbank, J. S. 75. Marshall, W. 75. Marshall, W. 76. Die Hyalonemen, Ein Beitrag z. Naturgesch. d. Spongien. Bonn, 1860. On the anatomy and physiology of the Spongi- adro. Ft. II.— Phil. Trans., clh. Supplement der Spongien des Adriatischen Meeres. Sur l'habitat du Hyalonema lusitanicum. — Proc. Zool. Soc. Lond. for 1865 ; p. 662. Notice sur deux espèces nouvelles d'Épongés de la famille des Lophospongiœ, — Arch, néerlan- daises sei. exact, et nat. T. III, p. 435. On the vitreous sponges. — Ann. & Mag. Nat- Hist., Ser. 4, I. Notes on deep-sea dredging. — Ann. & Mag. Nat. Hist., 4th ser., II, p. 423. Sur l'existence de la '' Holtenia Garpenteri " W. Thoms. dans les côtes du Portugal. — Jour. sei. math., phys. et nat. Lisbonne. No. IX, p. 69. On the ' Hexactinellidas,' or hexradiate spiculed siliceous sponges taken in the ' Noma ' Ex- pedition off the coast of Spain and Portugal. — Monthl. Microsc. Jour., Nov. 1870. Grundzüge einer Spongien-Fauna des Atlan- tischen Gebietes. Lpzg., 1870. On Holtenia, a genus of vitreous sponges. — Phil. Trans., clix. Fie Kalkschwämme. I. A monograph of the siliceo-librous sjionges. Ft. IV.— Froc. Zool. Soc. Lond., 1875. Untersuchungen über Hexactinelliden. — Zeit- schr. f. wiss. Zool., Bd. XXV, Suppl., [). 142. Ideen über die Verwand tschaftverhältnisse der Hexactinelliden. — Zeitschr. f. wiss. Zool., Bd. XXVII, p. 113. LITERATURE. 295 WiLLEMOES-SUHM, R. V. Schmidt, 0. Schulze, F. E. DönERLEIN, L. FlLHOL, H. Schulze, F. E. Schulze, F. E. Schulze, F. E. Agassiz, A. sollas, w. j. Lendenfeld, E. v. Topsent, E, Ijima, I. Wilson, H. V. Ijima, I. Schulze, F. E. '76. Von der Challenger-Expedition. Briefe an C. Tb. V. Siebold. VI!.— Zeitschr. f. wis?. Zool., Bd. XXVII, p. xcviii. '80. Die Spongien der Meerbusen von Mexiko (und des Caraibischen Meeres). II tes Heft. Jena, "1880. '80. On the structure and arrangements of the soft parts in Euplectella aspergilhim. — Trans. Roy. Soc. Edinburgh, XXIX, p. 661. '83. Faunistische Studien in Japan: Enoshima u. die Sagami Bai. — Arch. f. Naturgesch., Bd. 49, p. 102. '85. Ea vie au fond des mers. Paris, 188.5. '85. In : Narrative of the cruise of H. M. S. 'Challenger.' Vol. I. '86. Ueber den Bau imd das System der Hexacti- nelliden. — Abh. d. kgl. preuss. Akad. Wiss. Berlin, 1886. '87. Hexactinellida. Challenger Report. Vol. XXI. '88. Three cruises of the U. S. Coast and Greodetic Survey steamer 'Blake.' Vol. II, p. 171. '88. Tetractinellida. Challenger Report, vol. XXV. '92. Spongien der Adria. I. — Z. f. vv. Z. Bd. .53. '92. Contribution à l'étude des spongiares de l'Atlantique nord. In : Résultas des cam- pagnes scientifiques par Albert 1er, Prince souverain de Monaco. Fasc. II. '94. Notice of new Hexactinellida from Sagami Bay. I.— Zool. Anz., 1894, p. 365. '94. Observations on the gemmule and egg develop- ment of marine sponges. — Jour, of Morphol., vol. IX, no. 3. '95. On two new Hexactinellida from Sagami Bay. — Zoological Magazine (Tokyo), vol. VII, no. 79. '95. Hexactinelliden des indischen Oceanes. II Theil. Die Hexasterophora. — Abh. d. kgl. preuss. Akad. d. wiss. Berhn, 1895. 296 r. IJIMA : HEXACTINELLIDA. T. I.TIM A. I. Ijima, I. Maas, 0. topsent, e. Ijima, I. MiNCiiiN, E. A. Ijima, I. MiNCHIN, E. A. Moore, J. P. Schulze, F. E. Schulze, F. E. Schulze, F. E. Schulze, F. E. Takeshita, T. '93. Notice of new Hexactinellida from Sagami Bay.— Zool. Anz., 1896, p. 249. '96rt. Long-lines as zoological collecting apparatus. Zool. Mag. (Tokyo), vol. VIII. '36. Erlegte und strittige Fragen d. Schvvamment- wicklnng.— Biol. Centrallil., Bd. XVI, Nr. 6. '96. Résultats scientifiques de la campagne du ' Caudan' dans le Golfe de Gascogne. Éponges. — Ann. de l'Univ. de Lyon., fasc. II, p. 273. '97. Revision of Hexactinellids witli discoctasters, with descriptions of 5 new species. — Annot. Zool. Japonenses, vol. I, p. 43. '97. The position of sponges in the animal king- dom.— Science Progress, vol. I., n. s., July 1897. '98. The genera and species of Rosellidre. — Annota- tiones Zool. Jap., vol. II, part II. '98. Materials for a raonogra[)h of the Ascons. I. — Quart. Jour. Micr. Sei., vol. XL. '98. Hyalodcndron navalium, a new genus and species of Euplectellid sponge. — Proc. Acad. Nat. Sei. Philadelphia, 1898, p. 430. '99. Amerikanische Hexactinelliden nach dem Materiale der Albatross-Expedition. .lena, 1899. '99'''. Zur Histologie der Hexactinelliden. — Sitz. -her. d. kgl. preuss. Akad. d. Wiss. Berlin, 1899, p. 198. 19'. HexactinellideD^des indischen Oceanes. III Theil. — Abh, d. kgl. preuss. Akad. d. wiss. Berlin, 1900. 19'«. Die Jlexactinclliden. — In Fauna Arctica (Ergebnisse d. deutsch. Exp. in d. utirdl. Eismeer im Jahre 1898), Bd. I. 19'. In : Zool. Mag. (Döbutsugak-Zasshi, publ. in Tokyo), vol. Xrr, p. 261. LITEEATUKE. 297 N. B. — To uiy regret I have inadvertently failed to get access to the two following works : W. Maeshall. Spongiologische Beiträge. Festschr. z. 70sten Wiederkehr d. Geburtst. V. E. Leuckart. Leipzig, 1892. C. Gravier. Sur une collection d'épongés (Hexactinellides) du Japon. Bull. Mus. d'hist. nat. Paris. T. V, no. 8, pp. 419-é22. The former paper, as known to me through an abstract, should touch upon matters which demand attention in relation to the morpliology and physiology of the Hexactinellida. The latter paper may possibly contain notes on some of the species described in this Contribution. I beg indulgence for any omissions in reference to these works. — I also very much regret that I have not been able to benefit by Minchin's ' The Porifera ' (in Lankester's Treatise on Zoology), the ajjpearance of which work has become known to me too late to admit of my securing a copy of it before finishing the printing of this Contribution. 298 I. IJIMA : HEXACTINELLIDA. I. Contents. Page. Introduction 1 Topography of the Sagami Sea. (PI. XIV) 6 Collecting Hexactinellida and other deep-sea animals in the Sagami Sea 16 Methods used in the studies 32 EUPLECTELLID-ffi:. Eupledella Owen 37 Key to the species of Eupledella 58 Euplectella imperialis Ij. (Pis. I & II) 59 General characters 60 Spiculation 69 Young specimens 79 Miscellaneous notes. (Affinity of the species. Regeneration. Commensal immates) ... 83 Euplectella marshalli Ij. (Pis. Ill, IV & V) 86 General characters 88 Spiculation 96 Young specimens 103 Soft Parts 116 General arrangement of the soft parts and their relation tu the spicules ... 117 The dagellated chamber 128 The trabeculœ 147 Cells in connection with trabecular 165 * Archseocytes 165 Thesocytes 173 lieproductive elements 179 Suminaries of the histology of E. viarshidll 190 Development of hexasters 192 Miscellaneous notes. (Affinity. Regeneration. Commensal inmates) 200 Euplectella oweni Herkl. & Marsh. (PI. VI) .. 202 General cluuacters 204 Spiculation 211 Miscellaneous notes. (Abnormal sieve-plate. Commensal inmate, &c.) 215 Euplectella curvistellata n. sp 217 A\'.7af/;-e//a O. ScHM 220 Regadrella okiiioseana IJ. (Pis. VII & VIII) 223 General characters 224 Spiculation 229 Young specimens 239 Soft parts 247 Miscellaneous Notes, (llelation toi2. decora. Malformations.) 248 Regadrella komeyamai Ij- (Pis. IX à X) 252 General cliaractcrs 252 Spiculation 258 CONTENTS. 299 Regadrella phœnix O. Schm. (Pis. X & XI) General characters Spiculation Walterixt F. E. Sch Walteria leuckarti Ij. (Pis. XII & XIII) .. General characters Spiculation , Miscellaneous notes. (Monstrosity. Commensal Hydrozoa) Literature referred to in this Contribution Page. 265 .. 266 270 .. 278 279 .. 280 286 .. 292 294 PLATE I. Euplectella imperialis Ij. Plate I. EuplccteUa imperialis Lt. All figures in reduced scale of size. Fig. 1. An old specimen, 588 mm. long, of fully mature form. Fig. 2. A specimen, 491 mm. long, in which the up]ier part had not yet attained full size. Fig. 3. A rather 3'oung specimen, 245 mm. long. Jour Sei. Coll. Vol XV PI. I. EUPLECTELLA IMPERIALIS Ij PLATE IL Euplectella imperialis Ij. Plate II. Euplectelhi imperinlift Ij. Fig, 4. Upper end of a specimen, 465 mm. long., in wliich the wall is still growing in that part. Nat. size. Fig. 5. Portion of the upper end of a full-grown specimen, seen from the gastral side. Nat. size. Fig. 6. The youngest specimen in the Sei. Coll. Mas. Body 30 mm. long. Nat. size. Fig. 7. A young specimen, 72 mm, long. Nat. size. Fig. 8, Portion of the sieve-plate and cuflt', from a full-grown specimen. Reduced to -/^ nat. size. Fig. 9, Skeletal tube of a moderately large specimen, after washing away all the loose tissues. Nat. size. Fig. 10. Floricome at an early stage of developing terminals, 440 x , Fig. 11. A more advanced stage of same (sigmatocome stage). 440 x. Fig. 12. A still more advanced stage of same. 440 x . Fig. 13. Graphiocome in an early stage of developing terminals (rhaphides). 440 X. Fig. 14r<-f/. Different stages in the development of floricome-terminals, 1000 X . Fig. 15. Oxyhexaster. 440 x . Fig. 16. Lower end of a basal anchoring spicule. 100 x . Fig, 17. Oscularia in situ. To the right, the edge of the oscular membrane. 100 X Jour. Sei. Coll. Vol. XV. PI. II. EUPLECTELLA IMPERIALIS Ij. PLATE III. Euplectella marshalli Ij. Plate III. Euphcteïla marshalli Ij. Figs. 1-3, Three full-sized specimens iu about half natural size. Jour. Sei. Coll. Vol. XV. VI III. FX'PLECTELLA MARSHALLI Ij. PLATE IV. Euplectella marshalli Ij. Plate IV. Eupledella marshaUi Ij. Fig. 4. Gastral view of the wall. Sieve-plate at the upper end. Nat. size. Fig. 5. Showing the bottom-plate. Photo, from a drawing. Nat. size. Fig. 6. A young (22 mm. long), before the breaking through of parietal oscula. Nat. size. Fig. 7. Skeleton of a young specimen (32 mm. long). Nat. size. Fig. 8. A young specimen, 42 mm. long. Nat. size. Fig. 9. A young specimen, 63 mm. long. Nat ciz». Fig. 10. Floricome. 440 x . Figs. 11-13. Three stages in floricome development. 440 x. Fig. 14, Terminal of the immature floricome shown in fig. 13. With in- ceptional teeth. 1000 x . Fig. 15. Sieve-plates of Hwo young specimens (39 mm. & 48 mm. long) ; seen from above. 2 x . Fig. 16. Dermalia of an average size. 300 x . Fig. 17. Oxyhexaster. 400 x . Fig. 18. Young oxyhexaster. 440 x . Eig. 19. Portion of a graphiocome. 440 x . Fig. 20. Portion of a graphiocome after the los.-i of the terminals. 1000 x . Fig. 21. Lophocome. 440 x . Fig. 22. Space between the apopyles (op.) looked at from the excurrent s\dv. 440 X . c.vi., connecting membrane betw. the apoj)yles. lu.ia., marginal membrane of the chamber- wall, ir., trabeculje. Fig. 23. A portion of the dermal membrane. A trabecula exjianded into a film-like band. 1000 x . ir.n., trabecular nucleus, tli., an old thesocyte. Fig. 24. Portion of a thesocytal mass, showing the develoi>ment of thesocytes from archaiocytes (seen on the left). From a section stained with borax-carmine. 440 x . Fig. 2:"). Four thesocytes from above. 1000 x . Fig. 26, Lower end of a basal anclioring spicule. l.'iO x. Fig. 27. Oscularia in situ. Below, the oscular sdge. .50 X . Fig. 28. Section through the wall. 2;") x . — ^y.c/., arclucoeyte congeries, vh.L, chamber-layer, cx.tr., external trabecular layer, in.tr., internal trabecular layer, os.jii., oscular membrane, ose, oscularia. Jour. Sei. Coll. Vol. XV. PI. IV \^'^^,'^*v: 'S^-^-^^ 1 o.cl. ^■ J EUPLECTELLA MARSHALLI Ij. Lith Kushiba Ken da Tokyo Japan. PLATE V. Euplectella inarshalli Ij. Plate V. Euplectella marshalli Ij. Figs. 29, 30. Early stages in the development of floriconie. The principals inclosed in the scleroblast mass. Borax-carmine. 1000 x , Fig. 31. Portion of a mature floricome, still with the scleroblast-mass. Borax- carmine. 1000 X . Figs. 32-34. Stages in the 'development of graphiocome. Borax-carmine. 1000 X . Fig. 35. Optical section through the terminal sheaf of a developing graphio- come. Borax-carmine. 1000 x . Fig. 36. A small portion from the periphery in a section through the ledge, showing the conuli supported by dermal hilt-rays, the cobweb of trabecula3, &c. Borax-carmine. 300 x . — axL, archteocyte-congeries. arch., archfeocytes. eh., fundus of flagellated chamber, s.c, subder- mal cavity, th., thesocytes. Fig. 37. Chamber- wall (reticular membrane). Acid-fuchsin. 1,500 x. Above, the membrane is shown as out of the focus; the flagella are here visible either sidewise or in optical sections. The central dot in the choanocyte nuclei represents the origin of a flagellum. Below, at the right corner, four archreocytes (arch.) Fig. 38. Same with coarse refractive granules in the i)rotoplasm. Acid-1'uchsin. 1500 x . Fig. 39. Apopylar edge of the chamber-wall. Borax-carmine. 1500 x . — ch.n., choanocytal nucleus, wi.m., marginal membrane, ir., trabecula. tr. n., trabecular nucleus. Figs. 40, 41. Choanocytes in profile view. Acid-fuchsin. 1500 x. Fig. 42. Same in surface view, combined from views obtained at different foci of the microscope. Acid-fuchsin. 1500 x. Fig. 43. Optical section showing incurrent lacunar s])ace between four chambers. ]3orax-carmine. 1500 x . — i/i.r., membrana reticularis. Other letterings as in fio-s. 3 G & 39. Fig. 44. Groups of peculiar rod-like bodies (})robably nut belonging to tiie sponge). Borax-carmine. 1500 x . Fig. 45. Trabecula from a certain specimen, with egg-like cells (.r) of various sizes. Borax-carmine. 1000 x . — tr.n., trabecular nuclei. -/■ ?f^. 30. i^'H^ Jour. Sei. Coll. Vol. XV. PI. V -:t- 36. 37. arch •^ -^'«^ É^-^ -~^cLrch. ■^m^\si 33. \ 4\ «j^i; '® 38. ^4^. •'■■â'^'' '=<-'■ "f:;«'- ■?• •\ /• .*■«-'»«.* « '". «^.''^ ■ 'V ^ ''^i :^-A- I500X ,?5 9 i Ji i arch. '® J^. iiiiiiii ir.n. %-^ ^ i ra.r. ^r^''V. arch. \ :500x ' % "t tr. Ti . I600X ! , 1 1 • ! •• \ i ir. 1 ; ck.n. / ^ _ \ - ; > 1500« ~~ -^arch. ir.n. I500X .-• ♦'^ L arcK. mmi^ 'Tï^-'Si:- V ^ir.n 4S. '^^ EUPLECTELLA MARSHALLI Ij. l/th Koshiba Kanda Tokyo Japan. PLATE VI. Euplectella oweni Herkl. & Marsh. Plate VI. EuplectelJa oweni Herkl. & Marsh. Fig. 1. A specimen with comparatively prominently developed ledges. 7:5 nat. size. Fig. 2. Upper end of the largest specimen in the Sei, Coll. Mus. seen from the narrower side, -/s ^^^- '^i^^- Fig. .3. Lower portion of the same specimen, seen from the broader side. -/^ nat. size. Fig. 4. A specimen (Mr. Owston's) with little ledges or none at all. -/s nat. size. Figs. 5-6. Oxyhexasters. 440 x . Figs. 7-9. Anchor-heads of basal spicules. Figs. 7 & 8, abnormally de- veloped. Fig. 9, normal. 1.50 x . Fig. 10. Oscularia in situ. To the left, edge of the oscular membrane. 100 X . Jour. Sei. Coll. Vol. XV. PL VI l'IH.TOmo PRi«TIN8C.L= EUPLECTELLA OWENI Herkl. & Marsch PLATE VIL Eei^adrella okinoseana Tj. Plate VII. Regadrella olcinofieana Ij. Fig. 1. A complete and well-preserved specimen in the Sei. Coll. Mns. 72 nat. size. Fig. 2. A specimen without tlie upper end, bisected and seen from tlie gastral side, -/s nat. size. Fig. 3. Dead skeleton consisting of fused spicules. ^/.. nat, size. Fig. 4. Same of small size. Nat. size. Fig. 5. Portion of a dead skeleton, with 3 very young specimens of the same species attached to it. V/,, x . Figs. 6, 7. Two young specimens with still simple terminal osculum and with parietal oscula beginning to open through. Nat. size. Fig. 8. Pentactin-dermalia from a very young specimen, such as one of those shown in fig. 5. 1.50 x . Fig. 9. Immature floricome from a young specimen. 300 x . Fig. 10. Surface view of dermal layer from a very young specimen, con- sisting of pentactin-dermalia. Here and there, rhaphidial sheaves. 50 X. Fig. 11. Spiculation as seen in cross-section of the wall of a very young specimen with pentactin-dermalia. Above, the dermal surface ; graphiocomes, sheaves of rhaphides, &c. 50 x . Fig. 12. Spiculation of the wall (peripheral portion only) of a small speci- men, in which hexactin-dermalia have been added in large numbers to i^entactin-dermalia of an earlier developmental stage. 50 X . Jour. Sei. Coll. Vol. XV. PI. VII. REGADRELLA OKINOSEANA Ij. Lith Koshiba Kanda Tokyo Japan. PLATE VI 1 1. Eegadrella okinoseana Ij. Plate VIII. lîegadrclla okinoaeana Ij. Fig. ]3. Large liexactiii-dermalia (prostalia marginalia) from the cuff-edge. Ehapliides adhering to the distal ray. 150 x . Figs. 14-18. Hexactin-dermalia of varions sizes from the ledge. 150 x . Figs. 19-20. Oxyhexasters. 300 x . Fig. 21. Oxystauraster seen from side. 300 x . Fig. 22. Same seen flat on. 300 x . Fig. 23. Floricome. 300 x . Figs. 24-26. Intermedial microxyhexactins with spinose rays. 300 x . Figs. 27-28. Occasional intermedial spicules (canalaria ?). 150 x . Fig. 29, Section of the edge of a parietal ledge. 100 x . Fig. 30. Trabecultü, with thesocytes (ih.). 440 x . Fig. 31. Portion of sieve-plate beam, seen from dermal side. 50 x . Figs. 32-36. Some spicules shaken out from the dead skeleton shown in PI. Vir, fig. 3. — Fig. 32, spinose microxyhexactin (300 x ). Fig. 33, rare form of an intermedial hexactin (basidictyonalia ?) (150 x ). Fig. 34, basidictyonal hexactin (150 x ). Fig. 35, oxystauraster (300 X ). Fig. 36, remnant of a graphiocome after loss of terminals (300 X ). Fig. 37. Portion of oscular membrane, with spinose siüculcs which grade over into microxyhexactins. From a small and rather young specimen. 100 X. Jour. Sei. Coll. Vol. XV. PI. VIII. ^0 ?/: UTH. TOKIO PHINTINS CL» REGADRELLA OKINOSEANA Ij. PLATE IX. Eegad relia komeyamai Ij. Plate IX. Rcgadrella komeyamai Ij. Fig. 1. The type specimen in the Sei. Coll. Mas., consisting of two indivi- duals, -/ô ^^t- ^i^®' Fig. 2. Gastral side of the wall of the larger individual shown in fig. 1. Nat. size. Fig. 3. Cuff and corona of the same, seen from above. Nat. size. Fig. 4. Dermal surface of the same. Magnified about VJ.^ x . Jour. Sei. Coll. VoljgCV. PI. IX. F-^.» *^ä1.% *#- *» A .4 , -.«>-:i»*^. j REGADRELLA KOMEYAMAl Ij. PLATE X. Eegadrella komeyamai Ij. Eegadrella phœnix 0. Schm. Plate X. Figs. 5-17. lîegadrella homeyamai. Ij. Fig. 5. Portion cl' a floricome. 300 x . Fig. G. A terminal from the same. Lateral view. 670 x . Fig. 7. Terminal disc of floricome, seen from top. 670 x . Fig. 8. Large, unequal rayed, coronal oxypentactin. 5 x . Fig. 9. Portion of tlie free, coronal ray of the same. ÖO x . Fig. 10. Portion of the free ray of lateral prostal hexactins. 50 x . Fig. 11. Arrangement of prostal hexactins, dermalia, &c., at the edge of pleural prominence. 25 x . Fig. 12. Combination figure to show the spiculation of the wall. Above dermal, below gastral surface. 25 x . Fig. 13. Unusually large dermalia approaching a prostal hexactin. 150 x. Figs. 14, 15. Ordinary dermalia. 150 x . Fig. 16. Onychaster. 300 x. Fig. 17. A piece of the basal mass. Basidictyonal hexactins recognizable. 50 X . Figs. 18-27. Regadrella phœnix 0. Schm. (All figures from a specimen preserved on board the U.8. Fish-Commission Steamer * Albatross '). Fig. 18. Portion of a floricome. 300 x . Fig. 19. Same. 670 x . Fig. 20. Onychaster. 300 x . Fig. 21. Terminal claws of same. 1000 x . Fig. 22. Combination figure to show the spiculation of the wall. Above dermal, below gastral surface. 25 x . Figs. 23, 24. Ordinary dermalia. Figs. 25-27. Portions of large, sword-shaped dermalia found in groups on the lateral wall, around the hydranth of a commensal Hydrozoa. 100 X. Jour. Sei. C0II.J/0I. XV. PI. X. IITK.TOKIO PI1INT1N6 Ct?- 5—17 REGADRELLA KOxMEYAMAI Ij. 18— 27.?REGADRELLA PHŒNIX O. ScHM. PLATE XI. Eegadrella pliœnix 0. Scior. Plate XI. Regadrella pliœnix 0, Sohm. (All figures from a specimen on board the U.S. Fish-Commission Steamer ' Albatross '). Fio'. 1. Upper end of the specimen, seen from above. With damaged sieve- plate. Nat. size. Fig. 2. Upper })ortion of the. same in lateral view. Nat. size. Fig. .3. Small portion of dermal surface with a remnant of extremely deli- cate dermal latticework and with small groups of unusually large dermalia in association with the hydranth of commensal flydrozoa. About 10 X . Fig. 4. Large dermalia (prostalia marginalia) from the free edge of the ob- soletely developed cuff. 100 x . Figs. 5, G. Strong, unequal rayed oxystauractin and oxypentactin from the sieve-plate border, tlie superior ray of which gives support to radial beams of the sieve-plate. 5 x . Fig. 7. Small portion of the superior ray above-mentioned. 50 x . ' Fig. 8. Portion of a sieve-plate beam. 50 x . Jour. Sei. Coll. Vol. XV. PI. XI. Ç RKGADRKLLA PHCENIX O. Schm. Lith Kosh(ba Kan da Tokyo Japan. PLATE XII. Walteria leuckarti Ij. Plate XII. Waltcria leuckartî Ij. Ail figures in 74 ûat. size. Fig. 1. A well-preserved specimen which was in possession of Mr. Alan Owston. Upper end of the stem broken off. Fig. 2. A 8i)ecimen in the Sei. Coll. Mus. With all the tissues fallen off. Upper eud broken. Fig. 3. A well-preserved si)ecimen with the apical end intact. Lower part wanting. WALT K R I A LRU C K A R T I 1 PLATE XIII. Walteria leuckarti Ij. Plate Xin. Walteria leuckarti Ij. Fig. 4. The abnormally swollen, upper end of the stem in an otherwise normally developed specimen. -/:; "'^^- '^^^^'• Fig. 5, A part of the stem-wall cut open so as to show the gastral surface, -/s nat. size. Figs. 6-9. Sword-shaped dermalia. 150 x . Fig. 10. Pentactin-gastraha. 1.50 x . Eig. 11. Graph iocome. 300 x. Fig. 12. Spherical discohexaster. 300 x . Fig. 13. Same, partly broken. 300 x. Fig. 14. Portion of a terminal of same. To the right, terminal disc as seen from top. About 1200 x . Fig. 15. Onychaster. 300 x. Fig. 16. Principal and terminal ray of same. About 1200 x . Fig. 17. Another onychaster with more number of terminals. 300 x . Figs. 18, 19. Spinose parenchymal oxyhexactin. 300 x . Fig. 20. Terminal portion of a branch, magnified about 5 x . The numerous openings are those of cavities containing hydranths of the commensal Hydrozoa. At the lower end the cortical tissue has been stri[)])ed off, exposing the core composed of parenchymal diactins. Fig. 21. Cortical tissue and a portion of the core of a branch in longitudinal section. About .50 x . Above, the external surface with dermalia ; below, parenchymal diactins. — cav.liy., cavities containing the hydranth of the connuensal Ilydroza, with remnant of tissues belon-rinu: to the latter, cœn., cœnosarc of the Hydrozoa. I Jour. Sei. Coll. Vol. XV. PI. XIII. U /o J8 ■X.^ myf '/^ >-^N.^Lyra^ ^ miBB^^^maim^m WALTERIA LEUCKARTI Ij. Z/i/) Koshiba Kanda Tokyo Japgn. PLATE XIV. Topography of tlie Salami Sea. PI. XIV. Chart illustrating the -topography of the Sagami Sea. See pp. 6-15. Jour. Sc. Coll. Vol, XV, PI, XIV. Descriptions of Nine New Species of Fishes contained in Museums of Japan.^^^ By David Starr Jordan, Ph. D., LL.Ü. Pi csidcht, AND John Otterbein Snyder, A.M., lastrudor in Zoology, Ldand Stanford Junior University. {Communicated by Prof. 31itsuhuri.) With Plates XV-XVII. ■ 111 our recent investigations of the fishes of the Empire of Japan, several new species were observed each represented by specimens contained in museums of Japan, of which no dupli- cates were obtainable. The description of such species constitutes the purpose of the present paper. We would here express our special obligations to Professor Kakichi Mitsukurt of the Imperial University of Tokyo, and to Professor Chiyomatsu IsHiKAWA of the Imperial Museum at Ueno Park, Tokyo, for the privilege of examining and describing these species and for many other favors. The species here considered are the following : *The plates illustrating this article were prepared under my direction, and for whatever shortcomings they may possess, Dr. Jordan and Mr. Snyder are not responsible. — K. Mi- TSUKURI. 302 JORDAN & SNYDER : NINE NEW SPECIES OF 1. Acipenser kikuchii, Sagami Bay ; Imperial University. 2. Lepidopus aomori, Aomori ; type in Museum at Aomori. 3. Tetrapturus mitsukurii, Misaki ; Matsushima ; Otaru. 4. Tetrapturus mazara, Misaki. 5. Bentenia œsticola, Kashiina ; Imperial University. 6. Ebisus sagamhis, Misaki ; Imperial Museum. 7. Reinhardtius matsumurœ, Misaki ; no. 456, Imperial, Museum. 8. Trachypterus ishikaivcu, off Tokyo ; no. ;589, Imperial, Museum. 9. Trachypterus ijimce, off Tokyo ; no. 590, Imperial Museum. 1. Acipenser kikuchii Jordan and Snyder, new species. (PI. XV., figs. 1, 2). Head, 4 7.3 i'l body; depth 7; snout 2"/;? i'l head. Dorsal plates, 11 ; lateral 32 ; ventral, 11. Dorsal rays III, G3 ; anal, III, 37. Head longitudinally concave above ; snout shortish, rather sharp. Plates of back large, rugose or warty, with no distinct spines ; plates of side each with a spine in front ; those below smooth. Dorsal and anal each followed by a large rugose plate. No bony plates on body except a few small ones between the large anterior ones of dorsal series. Skin of body soft and smooth between the plates. Opercle rugose. Cheeks with fine stellate prickles. Height of dorsal contained 2 Ys times in head. Insertion of anal below posterior part of dorsal. Pectoral con- tained 17.1 in head. Caudal, from above, lYs times head. Described from a mounted specimen 180 centimeters long, in the museum of the Imperial University, Tokyo. Type locality, Misaki, Sagami Bay, Province of Sagami, Japan. FISHES CONTAINED IN MUSEUMS OF JAPAN. 303 The type ^Yas taken in the open sea in a net in deep water. The species is named in honor of Professor Dairoku KiKUCHi, the distinguished President of the Imperial University of Tokyo, in recognition of his interest in scientific research. This species is distinguished from most other sturgeons by the very long dorsal fin. From Acipenser mikadoi Hilgendorf, the only other Japanese species known, it is separated by the characters in the following analysis. a. Dorsal very long, of more than 60 rays ; anal of about 40; dorsal plates 11; skin between series of shields nearly smooth. Sagami Bay kikuchîi. aa. Dorsal moderate, of 35 to 40 rays ; anal of about 30 ; dorsal plates 7 or 8 ; skin between series of shields with small stellate plates. Rivers of Hokkaido (Ishikari River ; Streams of Teshio ; Mukawa) mikadoi. 2. Lepidopus aomori Joedan and Snyder, new species. Head 11 "/j in length ; depth 23 7^ ; dorsal spines 127 ; eye 5 Ya ill head ; snout 2 Yg. Maxillary not quite reaching eye ; 2 7g in head. Teeth moderate ; close set ; equal, except 4 strong canines in front of upper jaw. Pectoral 2 72 in head. No anal fin. Caudal very small, forked. Vertebrae 120. Color silvery. No ventral fins are evident on the dried skin. The dorsal spines are broken off; the above count being made from the neural spines at base of fin. Type, a dried specimen about 8 ft. long, in good condition except for the broken dorsal fin, preserved in the Museum of Aomori. Locality, Bay of Aomori, Province of Aomori, Japan. Local name, Ta chin u wo ; meaning sword-fish. 304 JOEDAN & SNYDER : NINE NEW SPECIES OF A similar dried specimen 6 ft. long from Hakodate is in the Fisheries Museum of that city. 4. Tetrapturus mitsukurii Jordan and Snyder, new species. (PL XVI, fig. 5).- D. XXXVII-6. Depth of body slightly less than height of dorsal iiu. Lower jaw from front of eye 74 more than postorbital part of head. Height of 1st dorsal equal to length of pectoral, the fin liigher than in T. mazara. Pectorals 1 "/s i^. head from tip of lower jaw ; ventrals 1 7i2' Caudal lobe as long as head from tip of lower jaw. Color steel blue with narrow whitish cross-bars on back ; dorsal violet, faintly spotted. Described from a specimen 6 ft. long without spear, examined in a fish-well at Misaki. Known to the fishermen as Makajiki or True Spear-fish. The species is generally common in Japan, but on account of its enormous size could not be preserved by us. It was seen at Misaki, Tokyo, Yokohama, Sendai and Otaru. The specimens from Matsushima Bay showed the following characters : D. XXXVIII-6 ; A. 14-7. Body slender, compressed ; its depth 4 in body exclusive of head. Lower jaw from front of eye a little more than post-orbital part of head. First dorsal low, equal to depth of body. Pectoral half longer than postorbital part of head or 1 Yj in head from tip of lower jaw. Ventral slightly longer than post-orbital part of head. Caudal lobe twice postorbital part of head. Color steel blue ; back above lateral line with about 15 * Kepiuchn'tion of the i)iK)tiigraiili refeneil to below. — K.M. FISHES CONTAINED IN MUSEUMS OF JAPAN. 305 vvliitisb cross-bars, faint and diffuse. First dorsal violet, vaguely spotted with black. These two specimens were each 8 ft. long, exclusive of spear. They were called by the fishermen Baisen or Kajiki- maguro. A large stuffed specimen from Otaru in the Fisheries Museum at Hakodate has the spear intact. Its length from eye is a little more than Y2 greater than rest of head. Tip of lower jaw a little nearer to eye than to tip of upper jaw : A photograph in the Imperial University taken from a specimen at Misaki shows the following characters : Head, with snout, 2 Ys iii length. Postorbital part of head 2 Ya in snout. Snout, from tips of lower jaw, 3 in head. Depth 8 V4 "^ length or 3 7s hi head. Dorsal 2 % in head, the length of its longest ray a little more than depth of body. Anal 4 in head ; pectoral 2 Vs > ventral "/ô ; lower lobe of caudal 2. The species is named for Dr. Kakichi Mitsukuei, senior Professor of Zoology in the Imperial University at Tokyo. 4. Tetrapturus mazara Joedan and Snydee, new species. D. XL-7. Back high, the profile rapidly rising to dorsal. Depth of body 1 Ys ill pectoral. Lower jaw from front of eye just equal to postorbital part of head. Spinous dorsal not abruptly falcate ; its height 1 7g in length of pectoral ; 1 Ys in length of body without head. Pectoral 1 Yg in head from tip of lowxr jaw. Ventral 1 ^5 in pectoral. Caudal lobe Yg longer than pectoral. Back dark blue with numerous whitish transverse bars ; both dorsal fins violet, with bright blue spots. 306 JOEDAN & SNYDER I NINE NEW SPECIES OF This species is much less common than the ordinary spear- fish or Kajiki. It is known to the fishermen as Mazara or as Kurokajiki (Black Spear-fish.). Our description is taken from the single specimen seen, 10 ft. long without spear, taken off Misaki, in Sagami. It is possible that either of these species may be identical with others described from other regions, but the evidence is against this supposition. T. indiens from Sumatra has never been intelligibly described. T. herscheli from South Africa is regarded by LÜTKEN as identical with T. brevirostris from India, a species which has longer ventrals. The Atlantic species, called T. imperato7', or 1\ helone, is close to T. miisukurii but differs in several regards. The rare T. amplus of Cuba is quite unlike either. The two species of Telrapiurus found in Japan may be thus distinguished : a. Pectoral fin moderate, 1 -/g in caudal lobe, 1 Ys ii^ head from tip of lower jaw ; dorsal lobe about equal to pectoral and about as high as body mitsuhurii. aa. Pectoral very long, scarcely shorter than caudal lobe and very little shorter than head, from tip of lower jaw ; back elevated at front of dorsal ; dorsal lobe shorter than pectoral and notably less than depth of body mazara. 5. Bentenia sesticola Joedan and Snyder, new genus und new species, (Flcraclidœ) ; (PI. XVL, Fig. 6). Head 4 in length ; depth 3 "/a ; eye 3 Vs ii^ bead ; snout 4; maxillary 2 "/-j', scales 49 ; D. about 55; A. about 40. Body elliptical ; closely compressed. Head with the profile before eye vertical. Mouth very oblique ; the lower jaw projecting. FISHES CONTAINED IN MUSEUMS OF JAPAN. 307 Maxillary broad, flat, scaly ; reaching beyond middle of pupil. Teeth fine, sharp, equal ; in narrow bands ; those on vomer and palatines similar. Anterior nostril round ; well separated from the posterior. No spines or serrce on head. Edge of preopercle membranous. All the bones of the head except lower jaw closely scaled. Preorbital moderate, sheathing the broad maxillary. Suborbital narrow. Branchiostegals 7. Gill- rakers 1 + 6 ; slender, small, far apart. Pseudobranchi^e large. Slit behind last gill moderate. Lateral line an ill defined streak. Scales hard and firm ; longitudinally striate ; those along dorsal and anal enlarged, papery, forming a deep sheathed groove into which the whole great fin fits and may be completely concealed. Scales on body bony, with oblique, angular, posterior edges ; those on lower parts each with a vertically compressed median spine. First dorsal spine at tip of nose, very short and slender ; 2nd, 3rd, and 4th progressively longer, yet slim and short ; the 5th, inserted over post- erior nostril ; very thick and very long, greatly enlarged (brok- en off in the type, but certainly more than twice length of the head). Rest of dorsal made up of slender, simj^le, inarticulate, flexible spines ; very long ; the anterior longest, reaching base of caudal ; the rest progressively shortened so that when laid back all end at about the same point ; the tips filamentous ; free from the thin, black, connecting membrane. (Whether the last few are semi-detached can not be clearly made out). Anal similar to dorsal ; the 2nd spine enlarged, half longer than head ; inserted just behind a vertical from eye. All the rays slender ; inarticulate ; technically spines. Vent directly below pupil. Pectoral long ; inserted 'ow ; slightly longer than head. Ven- tral minute jugular, 5 in eye ; the thin fragile rays almost obliterated and can not be exactly counted, probably 6 ; doubtless 308 JOEDAN & SNYDEE I NINE NEW SPECIES OF long ill young. Caudle on narrow peduncle ; moderately forked ; 1 Y2 in head. Color plain, metallic, lustrous, silvery; fins all black. Type, a large finely preserved alcoholic specimen about 18 in. long, in the Museum of the Imperial University. Locality, in the Kuro Shiwo or Japanese Warm Current off the coast of Kashima near Mito, Province of Hitachi. This beautiful species is allied to Pteraclis papilio Lowe from Madeira, also apparently a species of Benlenia, but it has more numerous fin rays. Bentenia^ is distinguished from Pteraclis vclifer and ccniropholis by the anterior insertion of its vertical fins, and by the enlargement of a spine in the dorsal and a spine in the anal fin. 6. Ebisiis sagamius Jordan and Snyder, new genus and new species, {Serranidce). (PI. XV., figs. 3, 4). Head 3 77 in length ; depth 3 -/s î snout 3 Yg in head ; maxillary 3; eye 6 % ; D. X-1, 13; A. Ill, 10; scales 122. Head large ; very convex in profile. Interorbital space very broad ; convexly elevated. Mouth rather small ; lower jaw pro- jecting ; maxillary extending nearly to middle of eye. Preorbital nearly as broad ns eye. Posterior nostril round. Preopercle entire. Opercle and other joarts of head without spines or serrations. Top of head with smooth skin ; sides scaly. Body covered with small scales which are loosely imbricated and quite rough. Lateral line normal ; not extending on caudal. Teeth strong; much larger than in Megaperca\ brush-like, in bands; no canines. Dorsals separate, both low ; dorsal spines rather weak ; anal with obscure spines (probably 3) ; longest ray 2 % in head. Pectoral broad and short, unsy m metrical, of 17 * Nanu'J for the Japanese goddess Kentex. — K. M. FISHES CONTAINED IN MUSEUMS OF JAPAN. 309 rays; its length contained 1 Ys i^^ head. Ventral inserted be- hind pectoral ; 2 -/s i" head. Caudal lunate ; sub-truncate ; 1 74 in head. Color dusky green, apparently clouded with darker. Type, a stuffed specimen 140 centimeters long, in the Im- perial Museum, Tokyo. Locality, Misaki, in Sagami. Known to fishermen as Abvrabözu, which means " fat-priest." It is also called Aburainagi or " Fat Bass." According to Kuma Aoki, an intelligent fisherman of Misaki, it is occasionally taken in the Kuro Siwo, it is not rare, and reaches a weight of 200 lbs. The genus Ebisus (named for the Japanese fisher-god Ebisu) is allied to Siereolepis and Megwperca, differing from both in the unarmed head, larger teeth and in the lower, weaker dorsal spines. 7. Reinhardtius matsuurae Jokdan and Snyder, new species. (PI. XVI, figs. 7, 8). Head 4 'U "^ length ; depth 3 V2 ; D. 96 ; A. 69. Scales 117. Body dextral. Interorbital width 3 in maxillary ; a little less than longitudinal diameter of lower eye ; cleft of mouth same on both sides. Lateral line single ; not sharply curved anywhere ; running obliquely downward to a point a little above middle of body and posterior to base of pectoral a distance equal to 2 times length of maxillary, then straight backwards to end of caudal fin, similar on blind side. Dorsal fin inserted just behind eye. Anal inserted below 26th dorsal ray. Dorsal and anal extending an equal distance posteriorly. Length of caudle peduncle 2 72 in head. Minute scales on interradial membranes of both dorsal and anal. Length of pectoral equal to maxillary. Color plain brown. A stuffed specimen about 1 7^ ft. long, no. 456, Imperial 310 JORDAN & SNYDER I NINE NEW SPECIES OP Museum, Tokyo. Locality, Misaki. This species is allied to Beinhardtius hippoglossoides, the Greenland Halibut, differing in the larger scales and in other characters. It is named for Mr. K. Matsuura, Curator of fishes in the Imperial Museum at Tokyo. 8. Trachypterus ishikawse Jordan and Snyder, new species. (PL XVII., fig. 10). Head 9 Y2 in length; eye 3 y.j in head; snout 2 7;,; maxillary 3 Vs ; D. 100. Preorbital very wide, radiate, rugose. Body gradually tapering backward, not constricted behind vent ; its depth about equal to length of head. Vent a little before middle of body, teeth Ö to 7 on each side of jaws ; the middle one longest. Lateral line running low along body ; its pores with spines. Ventral edge of body with tubercles throughout ; larger and rougher behind ; a number of hooked spines in pairs along lower part of tail ; body otherwise smooth. First spines of dorsal short and slender, not separated or elevated ; those near middle of fin much higher ; 1 ^4 in head. Pectorals 2 in head. Fin rays smooth. Color silvery throughout ; no spots. Described from a large specimen 1210 mm. long, nearly perfect but having the ventrals worn off showing only the basal bones at place of insertion, and the caudal lobe broken. The latter when entire probably measured 2 '/^ to 3 Vs in head. Type no. 589, Imperial Museum, Tokyo. Locality, off mouth of Tokyo Bay, between Misaki and Boshu. It is named for Dr. Chiyomatsu IsHiKAWA Curator of the Imperial Museum and Professor in the College of Agriculture in the Imperial University of Tokyo. FISHES CONTAINED IN MUSEUMS OF JAPAN. 311 9. Trachypterus ijimae, Jordan and Snyder, new species. (PI. XVII, fig. 9). D. VI -137. Profile vertical ; depth greatest at nape. Body abruptly constricted behind vent ; not gradually tapering as in T. ishikawœ. Eye larger, snout much shorter than in the latter. Lower part of tail with a double row of hooked spines ; 6 dorsal spines separated ; filamentous, their tips reaching past caudal. Ventrals little shorter than dorsal. Color silvery ; no dark spots. Type, a young specimen about 1 ft. long ; no. 590, Imperial Museum, Tokyo. Locality off the mouth of the Bay of Tokyo, between Misaki and Böshu. This pretty species is named for Dr. LsAO Ijima, Professor of Zoology in the Imperial University. Leland Stanford Jr. university. Sept. 20, 1900. PLATE XV. 1_ 10* Plate XV. Fig. 1 : — Acipenser hihucldi Jordan and Snyder. Side view. Pliotogi*a plied from the type-specimen. Fig. 2 : — Acipenser hikucMi Jordan and Snyder. Dorsal view. Photographed from the type-specimen Fig. 3 : — Ehisus sagamius Jordan and Snyder. Side view. Photographed from the type-specimen, -^q. Fig. 4 : — Ebisus sagamius Jordan and Snyder. Dorsal view. Photographed from the type-specimen, ^q i_ ID- Jour. Sei Coll. Vol. XV. PI. XV. k-'^, •^ K % PLATE XVI. Plate XVI. Fig. 5 : — Tetrapturus mitsiihurii Jordan and Snyder. Eeproduction of the photograph referred to in the text. Scale not ascertainable. Fig. G : — Bcntenia œstîcola Jordan and Snyder. Side view. Photographed from the type-specimen. |. Fis:. 7: — lîeinhardlius maisuurœ Jordan and Snyder. Dark side. Photographed from the type specimen. Fig. 8 : — Fiheinliardiius rnafsiturœ Jordan and Snyder. Blind side. Photographed from the type-specimen. Jour. Bei. Coll. Vol. XV. PI. XVI. PLATE XVII. Plate XVII. Y\g. 9 : — Trachypterus ijimœ Jordan and Snyder. Side view. Photographed from the type-specimen. Y\g. 10: — Trachypterus ishihawœ Jordan and Snydeh. Side view. Photographed from the type-specimen. 3_ '20" Jour. Sei. Coll. Vol. XV. Pl.XVII. Transpiration of Evergreen Trees in Winter. By Shunsuke Kusano, RigaTiusld. With Plate XVIII. I. Introductory. It is well known, from the researches of previous investi- gators, that evergreen trees in temperate climates can transpire even in the midst of winter. But as no such investigations have been undertaken with regard to plants indigenous to Japan, it has seemed desirable that efforts should be made to ascertain certain numerical values relating to the absorption of water by the roots of such plants and the evaporation of it from their leaves, during winter. Almost all the investigations with regard to transpiration, made up to the present time, by an enormous number of authors^^ 1) A fuller account with regard to transpiration is to be found in Burgerstein's excellent work, " Materialien zu einer Monographie betreffend die Erscheinungen der Trans- piration der rflanzcn." I, 1887 ; II, 1889. -^ 314 s. KUSANO : TRANSPIRATION OF have been confined to the vegetating season. That evergreen trees are constantly supplied with water even in winter, was first observed by Hales/' and then by DuhameF. Trevi- ranus"\ in his "Physiologie der Gewächse," says "In An- sehung der Jahreszeiten ist sie (Transpiration) unter gleichen Umständen im Frühjahre und Sommer am stärksten : im Herbste nimmt sie sehr ab und im Winter bemerkt man keine mehr." In the year 1860, T. Hartig^' made some experiments on trans- piration with Picea, a meter high, in milder winter, and found that the plant gave off from about 100 to 125 grams of water a day (V10-V4 Pfwnd) ; but as he gave neither the area nor weight of the transpiring part, we are unable to calculate the actual intensity of transpiration. Afterward Burgerstein^'^ pointed out in 1875, the relation of transpiration to lower temperatures and ascertained that transpiration of cut-branches of TaxiLS haccata was found to occur even in temperatures below zero. According to the results obtained by the latter botanist, Taxus haccata transpired in an hour at -2°C., 0.288 per cent, and even at -10.7°C., 0.019 per cent, of its fresh weight. A similar experiment was made by Wiesner and Paclier^'* with leafless cut-branches of Aesculus. In branches either one or three years old, the loss of w^ater at a temperature of-13°C. could still be observed. The above instances sufficiently prove that although the 1) Hales, Statik der Gewächse 1748, p. 29. 2)Duliamel, De l'exploitation des bois 17G4, Bd. I, p. 337. 3) Treviranus, Physiologie der Gcwiiclise 1835, I5d. I, p. 488. 4) T. Hartig, Ueberdie Bewegung des Saftes in den Ilolzpflanzen. Bot. Ztg., Bd. XIX, 1861, p. 17; and Lehrbuch für Förster 1877, 11 Aufl., Bd. I, p. 252. 5) Burger st ein, Ueber die Transpiration von Taxupzwcigcn bei niederen Temperaturen. Oesterr. Bot. Zeitschr., Bd. XXV, 1875. 6) Wiesner und Fächer, Ueber die Transpiration entlaubter Zweige und des Stammes der Rosskastanie. Oesterr. Bot. Zeitschr., lîd. XX\', 1875. EVERGKEEN TREES IN WINTER. 315 lower temperature affects the plant in diminishing the evaporation of water, it has but little influence in wholly stopping it. That the amount of transpiration greatly depends upon the temperature of the soil in which the plant grows, was first clearly shown by the well-known experiment of Sachs'^ He observed that some herbaceous pot- plants, e.g. Cucurbita, Nicotiana, &c., wither when the soil full of moisture was exposed to a temperature of 2-4°C., and he attributed this to the deficiency of the absorption of water. It must, however, not be concluded that, from the above experiments, the absorbing activity of the root of many plants in temperatures near the freezing point, or even below it, is completely destroyed : on the contrary, in several species of plants the root or even the cut-branches can absorb water considerably, as Kosaroff^^ has recently shown. The most interesting fact that the diminution of transpira- tion of evergreen trees in winter has a close relation to the closure of the stomata in that season, can be seen from the re- sults of the investigations made by several authors. Stahl, who laid stress especially upon this point, says: "Bei unseren immergrünen Straüchern und Bäumen, deren Existenz ohne den Spaltenschluss gar nicht möglich wäre, tritt derselbe schon frühzeitig im Herbste ein."''^ He has proved this fact by his " Kobaltprobe ""*'; and has shown that in some winter-green trees, for example, Hedera Helix, ten days were required to make the stomata reopen in a hot chamber. 1) Sachs, Das Erfrieren bei Temperaturen über 0°. Bot. Ztg., Bd. XVIIT, 1860, p. 124. Compare Sachs, Text Book of Botany 1882, 2nd Ed., p. 704. 2) Kosaroff, Einflu?s verschiedener äusseren Factoren auf die Wasseraufnahme der Pflanzen. Inaug. Disst. Leipzig. 1897. 3) Stahl, Einige Versuche über Transpiration und Assimilation. But. Ztg., Bd. LXX, 1894, p. 126. 4) 1. c, p. US. 316 s. KUSANO : TRANSPIRATION OF Subsequently, Lidforss^^ found that the guard-cells of the storaata on the leaves of some winter-green plants which he examined, were free of starch during winter, and that this absence of starch rendered the stomata incapable of performing their normal function. Although I have not made an extensive study of this jDoint, that is to say, to the extent of examining in each given case whether the stomata were surely closed or not, I have reason to conclude, so far as my observations extend, that many of our indigenous evergreen trees, unlike those of Germany above referred to, have their stomata more or less open even in the midst of winter. This condition may probably be considered as one of the chief causes which make the amount of the winter transpiration of our evergreen trees considerable. II. Method. The amount of water transpired by plants may be deter- mined in various ways : first, by weighing the plants themselves at definite intervals ; secondly, by condensing the vapour which is given off from the plants and measuring its volume ; thirdly, by measuring the increase of the weight of some hygroscopic substances, like calcium chloride, by which the vapour derived from the plant is absorbed ; and fourthly, by measuring the amount of water absorbed from the root or cut-surface. Of these four methods, only the first and the last were adopted in my investigations. ])Lidrorss, Zur I'liysiologic und Lldlogio dor Wiiitcigriincn Flora. But. Ct'iillbl., lÀl. LXVIII, 18%, p. 35. EVERGREEN TREES IN WINTER. 317 A. — 3Ieihod of determining the amount of water transpired by plants, by iveighing. The first method, recommended by many investigators as being the most accurate for experiments in which, of course, rooted plants must be employed, was fully discussed by Burgers te in^\ A number of evergreen trees, 40-60 cm. high, were selected for my experiments and planted last September in pots, measuring 15 cm. in diameter and 12 cm. in height. These pots in which the plants grew were enclosed within metallic cases of exactly the same form and size, and having bisected covers. For cementing the covers hermetically, I used tin-foil and a mixture of beeswax and olive oil. Through the covers a hole for supplying water was made which, however, was closed air-tight during the experiments. The whole weight of each pot, including the plant and the cover, amounted to about 2 kgr., when the soil contained in it was saturated with water, and I knew by calculation that about one third of the whole weight represented the quantity of water contained in the soil. By this method, I Avas able to make a rough estimate of the amount of water con- tained in the soil at different times during the experiments. Since the activity of the root is weakened in a closed soil owing to the deficiency of the air supply '\ experiments of long duration must be avoided. During the experiments I opened the hole in the cover several times, in order to supply water and also to renew the air. B. — IletJiod of oneasuriîig the amount of luater transpired by p)la7its, by absorption. The apparatus which I employed for measuring absorption was a slight modification of the potometer 1) Burgersteiii, Materialien zu einer Monographie betrellend die Erscheinungen der Transpiration der I'llanzen 1889, II, p. 5. 2) Sachs, Vorlesungen über PflanzenpliVbiologie 1882, p. 307. 318 s. KUSANO : TKANSPIEATION OF designed by MacDougal^^ One arm (Fig. 1 a) of a T-sliaped glass tube was bent parallel to the other arm ih) ; at the end of Fi^'- 1. nlllHlllllllMllnillllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllllMllll)ll I the former the branch was inserted, fitted with a rubber tube and bound with wire for safety. From one end of the latter arm, water was supplied by means of a stopcock, from the re- servoir, while the other end of the same arm was connected with the capillary tube which had an even inner diameter of nearly one millimeter. After the apparatus was filled with water, taking care not to leave any air bubbles in it, the loss of water absorbed by the cut-surface and transpired from the surface of the leaves was indicated by the diminution of the column of water in the capillary tube. Tlie volume of the T-tube must not be too large, since, if that be the case, a change of temperature gives rise to a change of the volume of water which will consequently IjMacDougal, A convenient potometer. Bot. Gazette, Vol. XXIV, 1897, p. 110. E VERGEBEN TREES IN WINTER. 319 cause an error in reading the water column. To prevent a violent change of the temperature of the water in the tube, the whole apparatus was immersed in a glass vessel filled with water, and the experiments were always commenced after a lapse of time sufficient to equalize the temperature both inside and outside the tube. By this means, I equated the temperature in the glass vessel and in the tube. The details obtained by this method will be given in the description of each special experiment. The amount of transpiration was reduced, for the sake of comparison with different kinds of plants, to the area of leaves in n dm., and also to their fresh weight and dry weight in 100 grams. For measuring the area of leaves, I employed the usual method of weighing pieces of paper cut in the same forms as the leaves, the weight of the unit area being previously ascertained. III. The Climate of Middle Japan. Before describing the details of my experiments, it will be worth while to give a brief account of the climate of middle Japan (Hondo) in which my observations were made. Since the climate of this island is greatly influenced by the ocean, it has a wide range and is rather inequable ; as a whole it is mild and highly favourable to a luxuriance of vegetation. Temperature. The mean temperatures during the winter in Tokyo'^ are 5.1°C. in December, 2.7°C. in January and 3.5°C. in February ; the average temperature being, therefore, 3.8°C. In January, the maximum is 15°C.; while the minimum is -6.5°C. 1) Calendar for 1899 published by the Imperial University of Tokyo. 320 s. KUSANO : transpieation of For the sake of comparison with certain localities in Europe, the names of the following cities with their respective mean tempera- tures in January are given, thus^-: — • Berlin 0.1°C. Modena 1.3°C. Munich -2.6 „ Florence 5.2 „ Vienna -1.2 „ Rome 6.7 ,, Triest 4.7 „ Milan 0.5 „ Hence the northern part of Italy is, in respect of tempera- ture, comparable with middle Japan. Humidity. Japan is comparatively wet during the winter, especially in the coastal regions, thus the relative humidity of the latter is'' : — Dec. Jan. Feb. Southern coast 73 9^ 69^ 79^^ Eastern coast 73,, 77 „ 76,, and also the relative humidity in Tokyo"^ is : — December Ç>5o/q January 65 „ February 67,, That tlie humidity in Japan is not so slight as to be injurious to plant«, as is the case during the dry season in tropical regions can also be ascertained by the following value of the rainfall and the number of rainy days in ïokyo^': — Dec. Jan. Feb. Rainfall 47.3 mm, 51.5 mm. 77.8 mm. Rainy days 6.2 6.7 9.2 and in Tokyo, the amount of rain tliroughout the year is 1463 mm. l)IIann, Handbuch der Klimatologie. Bd. Ill, 1S97. 2) Ko i de, Climatology ofjajian (in Japanese) 189S, p. 295. 3) Calendar for 1899 I.e. 4) 1. c. EVERGREEN TREES IN WINTER. 321 Temperature of the soil. The temperature of the soil is to be considered one of the most important factors affecting vege- tation. The number of days, when the minimum temperature of the earth's surface sinks below 0°C., is larger than that when the temperature of the air falls below 0°C.; thus the former is 91-121 days, while the latter 61-79 days. It is obvious that in winter an herbaceous plant rooted in shallow soil in the open ground, can not supply itself with a sufficient quantity of water for transpiration and the whole of it is destroyed, as we see in so-called annual plants ; but with regard to evergreen trees this is not the case, since their roots go deep into the soil where the temperature is not so low as to hinder the absorption of water. Thus the Central Meteorological Observatory gives the following observations^^: — Temperature of earth's surface. Dec. Jan. Feb. Mean 3.22°C. 3.C0°O. 4.08°C. Minimum 1.17 „ 1.37 „ 1.35 „ Maximmn 7.51 „ 8.24 „ 9.59 „ Temperature in soil, m. 0.05 deep 3.64 „ 4.00 „ 4.33 „ 0.1 „ 4.29 „ 4.26 „ 4.48 „ 0.2 „ 5.44 „ 4.80 „ 4.84 „ 0.3 „ 6.45 „ 5.18 „ 5.07 „ 0.6 „ 9.52 „ 7.51 „ 6.80 „ The warmer temperature, greater abundance of rainfall and higher humidity during winter months in middle Japan, in com- parison with the countries of central Europe, — Germany, for example — , lead us to anticipate that transpiration goes on much more actively in the former than in the latter. In this respect the northern part of Italy is perhaps in harmony with middle Japan. 1) Annual Report 1897. 322 s. KusANO : transpiration of IV. Evergreen Trees of Japan. The evergreen trees of Japan are numerous and luxuriant ; most of them being indigenous. Several kinds of Quereus and Lauracese, which form thick woods in southern Japan, are also found in the vicinity of Tokyo, where they attain a considerable height. Pasayiia {Quereus) cuspidata is one of the commonest evergreen trees more than 10 meters in height, and has densely foliate branches. Besides, we have Illicium Anisatum, Michelia compressa (Magnoliaceœ) ; Pittosporum Tobira (Pittosporaceœ) ; Photinia glabra, Erioboirya jaj)onieay Phaj^hiolepis japonica (Rosaceœ) ; some kinds of Citrus, SJcimmia japonica (Kutacece) ; Daphniphyllum macropodum (Euphorbiacea?) ; Ilex latlfolia, I. intégra, I. crenata (Aquifoliacese); Euonymusjap)onica (Celastracese); Tfiea japonica, T. Sasanqua, T. sinensis, Ternstrœmia japonica, Eurya ochnacea, E. japoîiica (TheacCcT) ; Daphne kiusiana (Thymekeaceaî) ; Eatsia japonica, Hedera Helix var. colchica (Araliaceœ) ; Aucuha japonica (Cornacefe) ; Aixlisia japonica (Myrsinaceas) ; Ligustrum japonicum, Osmanthus Aquifolium, 0. fragrans (Qleaceœ) ; and so forth. Most of them are shrubs or small trees, and in the vicinity of Tokyo are generally found in a state of cultivation. Species of Coniferte are also abundant. Pitins Thunbergii, P. densiflora ; Cryptomeria japonica ; ChanucGyparis obtusa, and Cephalotaxus drupacea have the widest distribution throughout Japan, extending from the southern to the northern parts. Podo- carpus Nageia, P. macrophylla ; Sciadopilys verticillata ; Juniperus 7'igida and J. sinensis are commonly found in the southern part ; and Abies fir ma ; Thujopsis dolabrata ; Thuja orientalis ; Torreya EVEKGKEEN TREES IN WINTER. 323 nucifera ; Fîmes parviflora ; Abies Veitchii ; Picea bicoloVy P. hondoensis ; Larix leptolepis ; Abies Mariesii, A. sachalinensis ; Picea ajaîiensis, P. Glehnii, etc., are found in the more northern part. Mayr^^ counted 14 groups with 30 species of Coniferse (Nadelholz) in all Japan, which shows their abundance. The leaves of these foliage trees are, with the exception of Araliaceœ, generally lanceolate and of a smaller size, with an entire or slightly serrated margin and a thick, hard and leathery texture. Almost all of them are hairless and have a glossy upper surface owing to the presence of a thick cuticular layer. As to their anatomical structures both cuticula and epidermal walls are tolerably well developed ; pallisade tissue generally consists of two (Quercus glauca, Ternslrœmia japonica, etc.) or three (Thea japonica, Pittosporum Tobira, Daphnifpliyllum macrojjodum, etc.) layers of cells compactly arranged. Intercellular spaces are diminished as is usually the case in xerophilous leaves ; while, the deep depression of the stomata in the epidermis is not to be found in the leaves of our evergreen trees. On the whole, it seems that our indigenous evergreen trees, in contradistinction to those in dry tropical or alpine regions, are less protected against transpiration. For my experiments, I selected, from among numerous species of trees belonging to different families, especially those whose anatomical structures differed the most widely. Experiments with cut-branches were made with materials found in the Botanical Garden. The species of plants used in my experiments were the following : — l)Mayr, Monographie der Abietineen des Japanischen Eeiches 1890. 324 s. KUSANO : transpiration op Coiiiferse — Oryptomeria jai^onica Don. Finns Thunhergii Pari. Fodocarpus sinensis Wall. ,, macrophylla Don. Torreya nucifera Sieb, et Zucc. Chamœcyparls ohtusa Sieb, et Zucc. Fagaceie — Quercus glauca Tliunb. Fasania cuspidata Oerst. Magnoliaceœ — Illicium Anisalum L. Berbericlacese. — Nandina domestica Thunb. Lauracese — Cmnamomum Loureirii Nees. Pittosporacese — Fittosporwn Tobira Ait. Kosaceaî — Eriobotrya japonica Lindl. Fhoiinia glabra Tliunb. Aquifoliacese — Ilex crenata Tliunb. Euphorbiacese — Daphniphyllum macropodum Miq. Theacese — Ternstorœmia japonica Tbunb. Thea japonica Nois. ,, Sasanqua Nois. Araliacese — Faisia japonica Decne. et Plancli. Cornaceai — Aucuba japonica Thunb. Oleaceœ — Ligusirum japonicw/i Thunb. ^Mh'vdcesß— Garde7iia florida L. Compositse — Ligularia Kœmpferi Sieb, et Zucc.'^' Liliaceie — Aspidistra elatior Bl.'^' Filices — Gynmogranwie japonica Desv.''' * Herbaceous plants. EVERGKEEN TEEES IN WINTER. 325 V. Transpiration under Direct Insolation. For this experiment I used the pots prepared as has been described above (p. 317). The pots were exposed all day to direct sunlight on a stand in front of the laboratory. To keep them free from rain, a glass roof was employed only during nights or rainy days, while in fine weather it was always put aside from morning till evening. Each pot was weighed once a day (4-5 p.m.), but I omitted the weighing several times, since the loss of water was too insignificant and the balance which I used was not sufficiently accurate under 0.5 gram. In the beginning of each series of experiments a sufficient amount of water was supplied to make the whole weight nearly 2 kilograms, at which weight the content of water might roughly be equalized in each pot. The weighing began at the end of December and lasted to the end of March, and in order to get a correct comparison at different times during the winter I noted, as far as possible, only the results of experiments in trans- piration obtained on fine days, thereby omitting those obtained on rainy or cloudy days. The materials employed were limited to the following fourteen species of plants, of which five species were conifers, and the others, foliage trees. Their characters and ages were as follows : — 326 s. KUSANO : TRANSPIRATION OF Name of plants. Ci'ijplom cria japonica Fimts Tliuriherj)iraliuii, but a.^ this uli.'ervation was made in bad weather it can not be considered to be normal. EVERGREEN TREES IN WINTER. 327 came about 3-6 times greater tlian at the end of January ; thus the average values of daily transpiration during January 17th- 24th and March 21st-23rd (see Tables III and VIII) were :— Daily transpiration jier D'!'"- in gr- End of January. End of Marcli. Quercus glauca 0,9öl 6.063 Pittosporum Tohira 0.506 2.012 lUicium Anisatum 0.462 1.974 Ternstrœm la japon ica 0. 32 8 1 . 802 Thea Jajmnca 0.331 0.934 Eribotrya japonica 0.476 2.006 Photinia glabra 0.395 1.140 Fatsia japonica 0.495 2.464 Dapjhniphyllum macrop)odum 0.434 1.251 As shown in the foregoing table, the minimum average value of transpiration lies between 0.328 (i.e. Ternstrœmia) and 0.50G gram (i.e. PiUosporum) per D dm. a day in the above nine species, with the single exception of Quercus (0.901). Of all the eight tables (Table I-VIII) we see that Quercus represents the maximum in the amount of transpiration, while Ternstrœmia shows, for the most part, the minimum. Other plants behaved themselves dif- ferently during different periods of the experiments. For the sake of comparison, therefore, I summed up the whole amount of trans- piration in each case, from the beginning to the end of the experiments^, and then reduced this to the unit area of leaves, as represented in the following table : — Names of Plants. . "^^^^^ amount of Trans- Reduced to tJie unit piration durin^; experiments. area of leaves O^'u. Quercus 345.0 gr. 62.1 gr. Pittosporum 402.5 „ 32.9,, llUcium 413.0 „ 27.5,, 20.4 gr- 22.2 36.2 28.7 35.7 26.4 328 s. KUSANO : transpiration op -yr e •D^ t Total amount of Trans- Redncerl to the tmit piration dnring experiments. area oi leaves Q t'ln- Ternstrœmia 142.0 gr. Thea 36.5.0 „ Eriohotrya 336.5 ,, Fholinia 327.5 , , Fatsia 386.0 „ Dapliniphylhim 386.5 ,, If these are arranged, according to the vahie of their amounts, in a descending order, they stand in the following succession : — QuercuSj Eriohotrya, Fatsia, Pittosporum, Photinia, Plicium, Daj)hniphylliim, Thea and Ternstrœmia. Äfode of transpiration. When we compare the intensity of transpiration of different kinds of plants, we see that their differences are smaller when the plants are in the period of their minimum transpiration, that is, at the end of January in the case of my experiments. If we take, for example, the amount of transpiration in the case of Ternstrœmia as a unit, the relative amounts in the other plants might stand as follows (conf. Tahle III) :- Qaerciis 2.75 FiUosporitm 1.54 lllicium 1.40 Ternstrœmia 1.00 Thea 1.00 Eriohotrya 1.45 Fhotinia 1.20 Faisia 1.50 Daphniphyllum 1.32 Thus, with the exception of Quercus, practically none exceeds one and a half times. EVEEGREEN TREES IN WINTER. 329 It is probable, though I have not made any accurate observations respecting it, that stomatal transpiration is, in this period at least, greatly checked ; but that a hermetic closure of the stomata does exist in these plants, as observed by Stahr^ and Lidforss"^ in most winter-green leaves, is doubtful judging from the results of the cobalt-test. If, on the other hand, we compare the transpiration in each plant observed in March, we obtain the following arrange- ment (conf. Table VIII) :— Querc7ts 3.92 Piftofiporum 2.1.5 lllicium 2.11 Teriistrœmîa 1.9.3 Thea 1.00 Eriohotrya 2.14 Photinia 1.22 Fatsia 2.63 Dapliniphyllum 1.34 Here the differences between them are greater, and the ratio ranges between 1 and 2.63 {Quercus being excepted). It may thus be seen that, in the coldest part of winter, the transpiration in various evergreen trees becomes approximate in amount, but diverges widely as the environment becomes favour- able to transpiration. The explanation of this phenomenon is not easily found, since the factors which act upon plants are complex ; but it is obvious that their influence varies with different species. This variability of transpiration becomes more apparent when the change of transpiration of each individual is traced 1) Stahl, I.e. 2) Lid for ss, I.e. 330 s. KUSANO : TEANSPIEATION OF through the different periods of time. To cite a few examples ; Thea transpired 0.6956 gram per n dm. per day at the beginning (conf. Table I), 0.331 at the minimal period^^ (conf. Table III), and 0.934 at the end of the experiment (conf. Table VIII). Their relation is, therefore, 2.1 : 1 : 2.8. In Ternstrœmia the amount of transpiration was 0.419 at the beginning, 0.328 at the minimal period and 1.802 at the end, so that their ratio Fig. 2. 2.1 2.0 1.9 1.8 1.6 1.5 1.4 l.P. 1.2 1.1 1.0 0.9 0.8 0.7 O.G 0.5 0.4 0.3 0.2 0.1 0.0 / / . / n \ / V / to 6 /i in the lower sides. In the first two plants, we have just the same number of stomata, viz., 300 per D mm ; while in the latter a smaller number, viz., 218 per D mm., but with greater dimension. As the variation in the anatomical characters of these plants is slight, so the amount of water trans- pired by them differs only a little (see Tables I-VIII). The remaining plants have rather thick epidermal walls ; thus in Plltosjyorum, the thickness is 14 fx in those of the ujoper side, and 9 1^ in the lower ; in Ternslrœmia 10 /i in the upper, and 10-5 /i in in the lower ; and in TJica 10-5.5 />« in the upper, and 10-4 /^ in the lower. In spite of the well developed epi- dermis in Pittosporum, we observe that the amount of trans- piration is far greater than in the case of Photiriia, Illiciuvi and Fatsia^ all of which have thinner cell-walls. This difference is most probably due to the larger number of stomata in the first named plant. A seemingly exce})tionul case is observed in Ternslrœmia, where in spite of the tolerably large size of the stomata and the EVEEGKEEN TREES IN WINTER. 333 abimdaiice of the intercellular spaces (Fig. 3), the amount of transpiration is very small, even less than that of FUtosporum which has smaller stomata and narrower intercellular spaces, al- though the number of the former is somewhat greater. This may probably be due to the checking of stomatal transpiration in winter, as may be seen by comparing the amount of transpiration at the end of March (vide Table VIII), when the amount suddenly in- creases, owing to the recovery of the function of the stomata. Fig. 3. Cross-section of a leaf of Ternstros)nia japonica showing the large intercellular spaces. X 97. The amount of transpiration in Thea is a little greater than in Terîisérœmia, but less than in all the other plants which I examined. Here, we see that all parts of the epidermal wall are thickened, the intercellular spaces become smaller, and the number of stomata amounts only to 293 per D mm.; all these characteristics point to the fact that this plant has well developed protection against transpiration in winter. Moreover, in Thea the increase Î34 s. KÜSANO : TRANSPIRATION OF of transpiration was not parallel with the increase of tempera- ture, as is shown by the curve in Fig. 2. My experiment wdth cobalt paper gave the following results for interpretation of this feebler transpiration : — On February 21st some leaves taken from a pot- plant of Thea were unable to redden the paper even after half an hour's exposure to direct sunlight, while at the same time Aucuba, Pittosjjorum, Photlnia, Ligustrum and Daphne odora all gave positive reactions, indicating open stomata. On April 23rd, the same experiment with Thea was repeated without obtaining any positive sign of storaatal transpiration. This limited value of transpiration may be attributed to a loss of function by the stomata, caused by the formation of thyloses (Thyllen) under the guard cells, to which Schwendener^^ first called attention, and of which I myself was able to find evi- dences in the leaves of Thea. The measurement of the epidermis and stomata are given as follows : — Names of plants. Quercus glauca Fatsia japonica Photlnia (jlahra Illic Iwii Anisatuiib Daplüiipliijllum mac7'opodnm, Tcrnstrœmia japonica Fittosporum Tobira Thea japonica Eriohoirya japonica Thickness of the outer- wall of epidermis. On the upper side of the leaves. 5/^ Ö 8 6.5 8 10 14 10-r).5 1,3 On the lower side of the leaves. 2.5^ 2.5 6.5 6.5 6 10-5 9 10-4 Number and dimensions of stomata.2) Number in in mm. 557 182 300 218 300 317 337 293 260 Leng III. 22^" 24 21 39 24 36 21 27 24 Breadth 1) ÖC h wendcnu r, Gcsaiuiiu'ltc Abhundkingeu Bd. I, IS'JS, 2) From Ikeno, I.e., excei)ling Quonus. p. 02. EVERGREEN TREES IN WINTER. 335 The intensity of transpiration in winter, as the results of my experiments show, seems to be so great as to indicate that the movement of water, or the corresponding activity of the root in absorbing water, still exists, even in winter, in a considerable degree ; and this fact becomes more obvious when we consider, on the one hand, the climate of Japan, and, on the other, the abundance of evergreen trees and also the general structures of their leaves (vide Chapter III and IV). Comparimn of transpiration between conifers and other ever- green trees. It was Höhn el who first compared the intensity of transpiration of conifers with that of foliage trees. By repeated investigations, he found that the intensity of transpiration of both kinds of plants stood at 1 : 6/^ I attempted to find out the difference of transpiration between conifers and other evergreen trees in Japan, and for this purpose parallel experiments with both kinds of plants were carried on during my investigations, the numerical data of which are given in Tables I-VIII. As the results of the above exjieriments, we found that the difference between them was very slight ; thus if we reduce the respective value shown in Tables III and VIII, for instance, to the fresh weight, and even to the dry weight, of the transpiring parts, and take its average in five species of conifers on one hand, and in nine species of other evergreen trees on the other, the relative amount of transpiration in them is roughly 1:2 or 1 : 1.5. Thus in Tables III, and VIII we find that the average amount of transpiration are as follows : — l)nölinel, Weitere Untcisnchungen über die Transpirationsjcrösse der forstliclicn Holzgewächse.— " Referat " in Just, Bot. Jahresbericht, Bd. VIII. 1, 1880, p. 241. 336 s. KUSANO : transpiration of Reduced from Table III. Reduced from Table VIII. Per cent, of Per cent, of Per cent, of Per cent, of fresh weight, dry wei.2;ht. fresh weight, dry weight. Cnifers 8.18 19.72 39.16 93.9 Foliage everOTeen trees 1G..58 37.74 64.G5 1.50.18 Katio 1:2.02 1:1.91 1:1.65 1:1.6 Altliough this slight difference exists in winter, there can be little doubt that in other seasons it would be greater, since, as we have explained in the foregoing paragraph (cf. mode of transpiration), the difference of transpiration in different plants is least during the cold winter but increases more and more when the outer conditions become more favourable. The feebler transpiration in conifers can be easily understood when we examine closely their anatomical characters ; the contrivances for protecting transpiration are highly developed especially in this class of plants. Their structures are characterized by the smallness of the transpiring surface, by the lignified cell- layer with its thick wall underlying the epidermis, by the deep depression of the stomata in the epidermis, and by the cuticular layer which attains a considerable thickness, etc.^^ Thus we see that in conifers, the development of xerophilous characters is perfect, while in other evergreen trees it is less so, and this anatomical difference chiefly causes the difference in the mode of transpiration in both kinds of plants. VI. Transpiration under Diffused Light. In order to understand the process of transpiration exhibited in successive sliort intervals, the method of absorption has been 1) Onnpare Thomas, Zur vergleichenden Anatomie der Coniferen-Laubhlättcr. Jahrb. f. wLss. Bot., Bd. IV, 18GG, p. 2:3. EVERGREEN TREES IN WINTER. 337 preferred here for ascertaining the amount of water transpired from cut-branches. A consideration of the relation between absorption and transpiration ought not to be neglected, when, as in the present instance, we use this method. While the outer conditions are constant, the relation between these functions also remains the same, but any violent change in the former at once modifies the latter ; as is shown by the experiments of many investigators.^^ Eberdt"' has also shown that the excess of ab- sorption occurs at night, while on the other hand, the excess of transpiration is observed in the daytime as the obvious result of the change of outer conditions, such as temperature and humidity, to which the plant is exposed during both day and night. To make the amounts of both absorption and trans- piration approximately equal, the external conditions, which effect these functions, must be kept constant, and in this case only can the amount of absorption be regarded as an indication of the amount of transpiration. A remarkable fact, not overlooked in my investigations, is that the absorption by cut-branches behaves differently from that of a rooted-plant. As the cut-branches are, strictly speaking, dying parts of the plant, not only is their absorbing power gradually weakened thereby, but also the filtrating activity of the cut-surface of the branches sometimes becomes weaker in conse- quence of the varying conditions to which they are subjected.'^^ It is requisite, therefore, that the experiments should be made with 1) Burgerstein, ]\Iaterialien. II, p. 55. 2) Eberdt, Die Transpiration der Pflanzen und ihre Abhängigkeit von äusseren Be- dingungen 1889. 3) Pfeffer, Pflanzenphysiologie 1897, 2 Aufl. Bd. I, p. 209. 338 s. KUSANO : transpikation of fresh cut-branches/^ Moreover in the case of cut-branches we meet with the existence of negative pressure, which may some- times produce great errors in the measurement of transpiration. Certain plants, for instance, Daphniphyllum macropodum, when cut off at 12. ,50 \).m. and observed at 1 p.m. and in each succeeding interval of 10 minutes, the following lengths of the column of water in the capillary tube were found to have been absorbed, under a constant temperature : — 226, 144, 120, 106, 102, 103, 89, 89, 88, 89, 88, 87, 85, 85, 82, 79, 79, 78, 79, 79, and so forth. In Pasania ciispidata taken at 8.50 a.m., I observed at 9 a.m. and every succeeding 10 minutes, as follows : — 47, 38, 31.5, 28.5, 26, 23.5, 23.5, 22, 22, 22, 21, 20.5, 20, 19.5, 18, 19, 20, and so on. The greater amount of absorption at the beginning of the experiments shows that the water is taken up by the plant in consequence of the existence of negative pressure, rather than merely to supply the loss of water effected by transpiration at that time. To avoid, as much as possible, such disturbing conditions, the experiments were carried on in a room where the temperature and humidity were kept nearly constant ; and the experiments were commenced two hours after the cut-branches had been placed in the room, when the negative pressure had nearly ceased and at the same time the branches had become somewhat accustomed to the conditions in the room. With regard to the preparation of the branch, a part of the required plant, the 1) A gradual diminution of absorption of water by cut-branches has been pointed out by Sachs (Flora, 1856). F. Darwin and W. Phillips have noted the precautions to be observed while using the potometer in their paper " On the Transpiration-Stream in Cut-Branclies " (Proceed, of the Cambridge Philos. Society, Vol. V, Pt. V, 1885, p. 330). EVEKGEEEN TREES IN WINTER. 339 cut-surface of wliich had been immediately immersed in water, was at first brought into the room and then a shoot vigorous enough for the experiment was cut off from it under water.^^ With tiiese precautions, I measured the intensity of trans- piration in some twenty kinds of plants, including, besides foliage trees, a conifer, a monocotyledon and also a fern ; and the results are given together as follows, denoting the quantity of water transpired from our evergreen trees in winter (Table IX ex- periment 1-20)'^: — per n dm. ] er hour, ingr. Gymnogramme japonlca 9 G.8 6 Quercus glauca 95.66 Thea Sasanqua 81.55 Ligidaria Kœmpfcri 71.71 Daphniphyll um macropodum 63.72 Thea japonlca 62.24 Eriohoirya japonica 59.15 Fatsia japonica 55.79 Pittosporum Tohira 55.36 Aiicuha japonica 54.64 Gardenia florida 53.54 Podocarpus inacrophylla 52.57 Nandina domestica 46.02 Pasania cuspidata 41.30 Cinnamomum LoureivH 40.95 Photinia glabra 32.53 Lig ustrum japonic urn 31 .53 Ternslrœmia japonica 30.56 Ilex crenata 24.62 Aspidistra elatior 6.48 1) Precautions for using cut-branches are given by Burgerstein, Materialien. II, p. 7. 2) The temperature and humidity were not constant in tlie different experiments ; the temperature varying from 11.5 to 7.4°C. in air, and from 12.6 to 6.4°C. in water. As to the details, reference sliould be made to that section in which the experimental data are treated of (Table IX). 340 s. KusANO : transpiration of Thus, among twenty kinds of plants, the intensity of trans- piration did not exceed 0.1 gram per D dm. per hour in diffused light. The greatest activity of transpiration was attained by Oymnogramme japonica, a species of fern ; it has thin herbaceous leaves, whose anatomical structure is characterized by the pallisade consisting of cells loosely arranged in one layer, and also by having an imperfectly developed cuticula. This fern, which grows in sheltered places, sheds its leaves in winter when it is found in open tracts. In such a slightly xerophilous plant, it is natural for us to expect a considerable loss of water. On the other hand, we see the least amount of transpiration in Aspidistra elatior, a monocotyledonous plant, whose cuticula is very thick, the mesophyll consisting almost entirely of compactly arranged parenchyma. Between these two extremes of transpiration stand the typical evergreen trees, ivhich emit, on an average, 53 mgr. of water per ^ dm. per hour. Among these evergreen trees again, the maximum amount is attained in Quercus glauca, while the minimum amount is found in Ilex crenata. We see in our experiments, that the relative amount of transpiration in pot-plants on one hand, and in cut-branches on the other, do not correspond with each other, as for instance, tlie amount transpired by the cut-branches of Thea japonica was greater than that of Pittosporum ; while in the case of potted plants, the former was surpassed by the latter. This diversity in the amount of water transpired in the two different cases is partly due to the difference of conditions (temperature, humidity, light &c.) under which the experiments were made ; but chiefly to the methods of the experiments, for in one case the entire plant was used, while in the other only a part was employed. Moreover the difference between individual plants, and also between EVEEGKEEN TREES IN WINTER. 341 different parts of the same plant is an important factor causing a great diversity of results in a given experiment ; as we see in the researches of Kröber^^ who found, for instance, that the difference in the amount of transpiration between any two branches taken from the same plant was sometimes greater than the difference between two branches taken from two different plants of the same species. A difference between different plants was observed by me in Thea japonica : some individuals of this plant which I examined either in pot or in garden showed closure of stomata after the " Kobaltprobe," while others had the apertures com- pletely opened at the same time. It is, already, well known that a considerable amount of water is given off even in a temperature near 0°C. I witnessed the same fact in the cut-branches of a few evergreen trees. Thus for example (Table IX experiment 21-27) : — The amount of transpiration per D dm. per hour, mgr. Thea japonica 25.71 Ternstrœmîa japonica 24. 84 Dapliniphyllum onacropodum 22.87 21.07 Fiitosporum Tohira 19.99 Aucuba japonica 13. GO 13.69 Pasania glabra 9.90 Averaççe amount 18.96 Comparing these results with those of the preceding ex- periments (expt. 1-20), we see that the effect of low temperature upon transpiration seems to vary according to the nature of the plants. Thus in Aucuba japonica, the intensity of transpiration 1) Kröber, 1st die Transpirationsgrösse der Pflanze ein Maasstab für ihre Anbau- fähigkeit? Landwirtsch. Versuchst., Bd. XXIV, 1895, p. 503. 342 s. KUSANO : tkanspieation of ill the latter cases (expts. 22 and 26) is reduced to one fourth of that in former case (expt. 1), and in Daphniphyllum macro- podum as well as in PUtosporum Tobira, to one third ; while Ternstrœmia japonica transpires very sluggishly without showing any noticiable difference during the experiments in both high and low temperatures. This slight variation in the quantity of transpiration of Ternstrœmia japonica under such condition of temperature was ceteris paribus a noteworthy phenomenon during these experiments ; thus again, the amount of transpiration, under air temperature of 9.2-9.°C. and water temperature of 9.2-9.6°C. in December was 30.56 mgr. per D dm. per hour, while at the lower temperature, both of air and water, of 2.6-2.8 and 1.5°C., 24.84 mgr. of water was given off. The cause of this slight difference becomes evident when we consider the presence or absence of stomatal transpiration. Lastly let us describe here, for the sake of comparison, some experiments made, in the midst of summer, in a room under diffused daylight. At that time the weather was very wet and almost rainy, the relative humidity indicated being 80-909^, or even more. Helianthus tuberosus with 56 leaves, standing in a glass vessel filled with water, transpired 2.511 grams per D dm. in the interval between 12 m. and 6 p.m. on the 25tli of August, and 3.349 grams per D dm. between 6.20 a.m. and 5.20 p.m. on the next day. The amount of water transpired by the space of a square decimeter of the leaves in an hour was, therefore, 0.418 gr., in the first, and 0.304 gr., in the second experiment. Thus with the typical evergreen leaves observed in the cold winter on the one hand (at a temperature near 0°C., — Table IX experiment 21-27) and with the typical summergreen leaves observed in the midst of summer on the other hand, the relative difference of trans- EVERGREEN TREES IN WINTER. 343 piraiion between two serîes of observations was great, and can be indicated at about 1 : 20. Recently, Kosaroff in his dissertation, quoted above, has experimentally shown that rooted plants as well as cut-branches can absorb water even under 0°C.; and also that in the stem of trees water can pass through when a portion of the stem is cooled below 0°C. My experiment has been carried on with the view of ascertaining the mode of absorption of water by plants when they are exposed to extreme cold. Such conditions com- monly occur in plants on cold winter morning when the leaves are sometimes frozen stiff. As is well known, the absorption of water by cut-branches, in ordinary air temperature, is greatest at first in consequence of the negative pressure of gas, and gradu- ally diminishes as the pressure comes to an equilibrium. How- ever, when the branches were cut off on a cold winter morning and were brought into the room, a quite different phenomenon was found to occur ; the power of absorption was at first very weak but after a short time, it suddenly increased and then sank gradually until it became constant in each succeeding interval. As it appeared to me that this fact might have a close connection with the process of transpiration on a cold morning, I endeavoured to examine it more closely. In the middle of last winter I cut off some shoots and im- mediately measured the amount of water absorbed by them in a room in which a nearly constant temperature above 0°C. was maintained. In Daphniphyllum macropodum and Aucuba japonica, the alteration of the absorbing power was remarkable. In the former plant (see Table X) 8 mm. of water column was absorbed during the first hour, but after one hour 245 mm. was absorbed 344 s. KusANO : transpiration of in only 10 minutes (compare Curve X in Plate XVIII). In the latter plant, the absorption of water in the first 10 minutes was 5 mm. but after about one hour it reached the maximum of 50 mm. (vide Table XII and Curve XII). This mode of absorption dej)end3 not only upon either the temperature of the open air or that of the room, but also varies according to tlie nature of the plants used in the experiments. In a branch of Thea japonica, cut off on a very cold morning and brought into a room, in which the temperature stood at 1.8°C., and l.O'^C. in air and water respectively, I could not observe any increase of the power of absorption, but, on the contrary, when the cutting surface was made, the power was most visforous at first which we often observe to be the case in an ordinary temperature. In Aucuha japonica (vide Tables XII and XIII) and TernstTœmia japonica (vide Table XIV), a retardation of the absorption was indicated even at a little higher tempera- ture. In order to reach the maximum degree of absorption, more than one hour after the branch was brought in the room, seemed to be necessary in my experiments ; thus, in Dap>hniphyllum, it was reached in one case after two and a half hours, and in another after two hours and forty minutes ; in Aucuha the interval was in one instance one hour and twenty minutes and in another nearly two hours ; in Pamnia glabra, one hour and twenty minutes ; and in Ternslrœmia, only forty minutes (compare Table X-XV and Plate XVIII). This rapid absorption of water during the first few hours is due not to the suddenrise of transpiration, but to the restoration of the absorping power itself. This may take place in the follow- ing way : — Exposure of the plants during the night to a cold EVERGREEN TREES IN WINTER. 345 temperature many degrees below 0°C. rendered the absorption of water more difficult. Owing to the deficiency of water the plant lost their turgidity of tissue and consequently became wilted. As the temperature after daybreak gradually increased and reached about 0°C. or above, the absorption of water became gradually easier, until after about an hour, it attained its maximum. When a sufficient quantity of water was thus taken up, the rate of absorption became slower and constant, and the wilted branches gradually resumed their norm^al position and stood perfectly erect. VII Summary. The results of my investigations may be briefly summarized as follows : — 1. The evergreen trees indigenous to Japan used in my experiments transpired in winter in Tokyo, an average quantity of, at least, 0.48 gr. per D dm. per day (with the exception of conifers), or 16.58 gr. per 100 grams of fresh weight in foliage trees, and 8.18 gr. in conifers per day. 2. In the southern part of our country where the climate is milder (the mean temperature at Nagasaki in January being 5.°C.), the intensity of transpiration would undoubtedly be greater. But the contrary is no doubt the case in the northern part, especially in the island of Yezo, where the winter is severe (the mean temperature at Sapporo in January being -6.3°C.), and the plants must protect themselves from a great loss of water ; and perhaps we may in their case expect the same occurrence of a minimum transpiration, as has been observed, for example, in Germany, 346 s. KUSANO : transpiration of 3. Not only is the transpiration continued in winter in Tokyo, but also the assimilation, as Miyake^^ has recently shown, takes place without intermission in winter, though it is much feebler than in summer ; and the non-cessation of these principal physiological functions in winter would naturally lead us to conclude that the abundance of evergreen trees in Japan is chiefly due to the favourable climate. 4. The time of minimum transpiration agrees with that of the minimum temperature, and occurs at the end of January. 5. The difference in the amount of transpiration in different species of evergreen trees becomes smallest at the time of mini- mum transpiration ; and a change in the external conditions, especially in temperature, does not necessarily produce a corres- ponding change in transpiration in different species. 6. In average cases the amount of Avater transpired by foliage evergreen trees, is one and a half or two times greater than that transpired by conifers if we reduce the amount either to the fresh weight or to the dry weight of the transpiring part. 7. In diffused light at a temperature of ca. 10°C., the average transpiration of many evergreen trees amounts to 53 mgr. per □ dm. per hour. The present work was undertaken, at the suggestion of Prof. M. Miyoshi, during the academic year of 1898-1899, and, under his direction, I was able to carry on a large number of experiments. To both Prof. J. Matsumura and Prof. M. 1) Bot. Centlbl. Bd. LXXX, 1899, p. 172. EVERGREEN TREES IN WINTER. . 347 Miyoslii I wish to offer my heartiest thanks for their kind advice during the progress of my work in the laboratory of the Botanical Institute belonging to the College of Science. I am also indebted to Mr. K. Nakamura, Director of the Central Meteorological Observatory in Tokyo, for his kind per- mission to use the climatological tables made in the Observatory there, and to all other friends who have kindly assisted me in various ways. July 1899. 348 s. KusANO : transpiration of ERRATA. Page 327, line 1, far 3-6 times, raid 3-5 times. „ „ „ 6, for 6.063, read 3.661. EVERGREEN TREES IN WINTER. 349 EXPERIMENTAL DATA. I. Transpiration by Direct Insolation of Pot-Plants. With regard to the temperature and relative humidity which are referred to in the following experiments, I adopted the observations made by the Centra] Meteorological Observatory of Tokyo, which is situated at a distance of one and a half miles from our laboratory. Both places are almost similar not only in position but also environment, and the hourly observation shows that the air temperature in both places are nearly the same, as will be seen from the following comparison. Sept. 27th. At Bot. Gard. At Obs. Sept. 29tli. At Bot. Gard. At Obs. 3p.m 23°C 23.6°C. 10 p.m 16.2 16.4 4 22.5 22.4 11 16.2 16.3 12a.m 16.2 16.3 28th. 1 p.m 16.25 16.3 lOa.m 21.5 21.5 2 16.5 16.5 11 22.75 22.3 3 15.75 15.9 12 23.5 22.8 30th. 1 p.m 23.5 23.9 7 a.m. 17.5 17.8 2 23.5 24.6 8 20+ 20.1 3 23.0 23.5 9 21.5 21.8 4 21.75 21.8 10 23.5 23.4 5 20.0 20.2 11 24.75 24.1 In the column representing the total amount of transpiration in the following tables, I have given the total daily transpiration ; and in the column of average of daily amount of transpiration, the mean value for each day. In the succeeding three columns, I have given the value of trans- piration, calculated from the average of the daily amount of transpiration and reduced to the area, fresh weight and dry weight, of the transpiring parts respectively. The amounts were expressed all in gram. \50 s. KUSANO : TRANSPIEATION OF TABLE I (December 28, 1898-January 4, 1899).* Weighing was made at the beginning and at the end ; during the interval the glass cover was not removed. The weather was fine throughout the day. Names of plants. Oryptomeria ... Pimis Podocarpus ... Torreya Chamcecyparis Quercus Pittosporum ... nudum Ternatrœmia ... 2'hca Enoholrya Photinia Fatsla Daphniphyllum Total amount Average of daily amount of transpiration. 95.0 91.5 65.5 58.0 39.0 68.0 73.5 81.0 20.5 80.0 64.5 70.5 67.0 38.0 of transpiration. 13.571 13.071 9.357 8.285 5.571 9.714 10.500 11.570 2.928 11:418 9.210 10.071 9.571 11.714 Transpiration during 24 hours. per D dm. of surface. 1.742 0.833 0.769 0.419 0.6956 0.990 0.883 0.885 0.800 per 100 gr. of fresh weight. 18.4 12.3 8.3 21.2 14.8 77.2 26.9 24.9 12.6 22.9 30.3 32.4 21.7 32.8 per 100 gr. of dry weight. 38.6 33./ 18.7 55.6 30.9 161.1 70.1 48.7 32.5 35.9 63.6 70.3 74.8 78.6 29th 30th 31st 1st 2nd 3rd 4th. /-Mean G.O 5.2 3.7 2,9 3.3 5.7 6.0 • Temiieialiirc^Maxiiuuiu ...11.7 ill 9.3 10.8 9.5 10.6 11.4 iMiuimum ... 1.9 1.8 -1.1 -2.2 0.4 1.3 -0.6 Kelatitve humidity 71.9 41.3 47.7 66.7 72.2 68.2 71.8 EVERGREEN TREES IN WINTER. 351 TABLE II (January 4-11). During daytime the glass cover was put aside and the weighing was made every evening at 4 p.m. The weather was every fine throughout the day. Names of plants. Daily amount of transpiration. Date. £^ Mea. S Max. H Will. Humid. 4-5 -1.0 63.4 6 3.1 7.8 -1.5 49.2 -1.2 57 9 8 3.3 12.4 -1.5 65.6 3.5 9.9 -3.0 71.3 10 5.5 11.9 2.0 11 5.7 10.9 1.5 58.4 2 u Transpiration during 24 hours. -=0.2 O C3 cS o a I en O +^ o & . J •- ;;; so •- g ^ 9.286 — 12.6 26.4 5.714 — 5.4 14.7 7.929 — 7.0 15.9 6.642 — 17.0 44.6 4.714 — 12.5 26.2 7.429 1.34 59.1 123.2 9.214 0.731 23.6 61.6 9.071 0.604 19.6 38.2 3.071 0.440 13.2 34.1 10.857 0.662 21.8 31.4 6.643 0.714 22.3 45.8 7.286 0.639 2.3.7 50.9 7.64:} 0.701 17.3 59.8 8.000 0.546 22.4 53.7 Oryptomeria Pi7lUS Podocarpus Tori-eya Chamœcyparis ... Quercus Pittospoi-um HUcium Temsirœmia Thm Eriboirya Photinia Fatsia Daphniphyllum... 11.5 10.0 10.0 8.5 7.0 8.0 8.5 9.0 4.0 14.0 8.0 8.0 9.5 12.0 8.5 5.5 8.5 7.5 4.5 8.5 11.5 12.0 2.5 15.5 8.0 8.0 8.5 8.0 9.0 6.5 7.0 5.5 5.0 8.0 7.5 8.5 3.0 12.0 7.0 6.0 8.5 8.0 8.5 6.5 9.0 12.0 65.0 3.5 3.5 4.5 6.5 40.0 8.0 5.0 7.0 10.0 55.5 6.0 5.5 5.5 8.0 46.5 3.5 3.5 3.5 6.0 33.0 5.5 5.0 6.0 11.0 52.0 7.0 6.0 8.0 16.0 64.5 8.0 6.5 8.5 11.0 63.5 3.5 3.0 2.0 3.5 21.5 9.5 5.5 8.0 11.5 76.0 3.0 3.5 7.0 10.0 46.5 6.5 4.0 7.0 11.5 51.0 6.0 3.0 8.5 9.5 53.5 6.5 6.0 7.5 8.0 56.0 352 s. KUSANO : TRANSPIEATION OF TABLE III (January 17-24). During daytime the glass cover was removed and the weighings were made every day at 4 p.m. The weather was very fine throughout tlie day. Names of plants. Daily amount of transjii ration. Date. c, Mea. Ï Max. =^ Min. 17-18 2.9 8.9 -3.3 39.2 2.5 10.5 -4.2 5C.C 20 21 22 23 2.8 1.4 1.9 0.7 7.7 6.3 8,9 8.9 -].S -2.G -3.1 -5.9 61.5 48.8 47.7 74.2 24 3.1 7.5 -0.8 72.3 'S 0.2 Transpiration during 24 hours. o ?i Oryptonieria Pinus Podocarpus Torreya Chamoecypans ... Quercus PUiospoi'um niicium Ternslramia Thea Erioholrya Photinia Fahia Daphniphylluni . . . 10.0 10.0 5.5 9.0 2.0 5.5 6.0 9.0 2.0 6.0 5.5 4.5 7.5 6.5 7.5 7.5 7.0 9.0 5.5 6.0 10.5 11.0 5.5 3.0 7.0 5.5 4.0 7.5 3.5 5.5 5.5 6.5 7.0 4.0 3.0 2.0 2.0 2.0 1.0 4.0 6.5 5.0 7.5 2.5 7.5 6.5 7.5 6.0 4.5 6.0 8.5 8.0 7.0 0.5 2.5 2.5 2.5 3.5t 1.5 3.5 6.5 6.5 7.5 5.0 4.0 5.5 5.5 5.5 2.0 5.5 5.0 0.0 4.5 4.0 7.0 5.0 5.5 5.5 4.0 6.5 6.5 7.5 8.0 6.0 4.5 4.0 2.0 2.5 2.0 4.0 2.5 3.5 1.5 3.0 3.0 2.0 3.0 3.5 51.0 50.0 35.0 40.0 14.0 35.0 40.5 48.5 16.0 38.0 31.0 31.5 37.5 44.5 7.286 7.143 5.000 5.714 2.000 5.000 5.786 6.929 2.286 5.429 4.429 4.500 5.357 6.357 0.901 0.506 0.462 0.328 0.331 0.476 0.395 0.495 0.434 9.9 6.7 4.4 14.6 5.3 39.7 14.8 14.9 9.8 10.9 14.7 14.6 12.1 17.7 20.7 18.4 10.0 38.4 11.1 82.9 38.7 28.1 27.7 15.7 30.6 31.4 41.9 42.7 t A leaf liad fallen off. evehgeeen trees in winter. 35^ TABLE IV (January 24-28). Between the preceding night and the forenoon of the 25th there was a snow fiill ; the 28th was cloudy, but all the other days were fine. The glass cover was removed during the daytime. Weighings were made at 4 p.m. every day. Names of plants. Daily amount of transpiration. Date. j£, Mea. S Max. H Min. Humid. 24-25 0.8 5.5 -1.7 S5.9 26 1.5 7.5 27 2.9 9.8 -4. 54.7 28 1.7 4.9 -2.5 a c 0 cS d c 's c 1 a ii a t- rt rt Transpiration during 24 hours. S o e: Crypiomeria Pinus Podocarpus Torreya Chamœcyparis ... Quereus Piltosporum Tllicium Ternstrœmia Thea Erioboirya Photinia Fatsia Daphniphyllum . . . 6.0 7.0 6.5 7.5 3.0 4.5 2.5 7.0 2.0 1.0 3.5 5.5 3.5 6.0 3.5 7.0 1.5 2.0 3.0 6.0 4.0 5.0 3.5 5.0 4.0 6.5 7.5 7.0 8.0 3.5 6.0 5.5 1.5 5.5 6.0 7.0 3.0 6.0 4.0 5.5 55 7.0 3.5 24.5 6.125 3.0 20.5 5.125 2.5 16.0 4.000 2.5 17.5 4.375 0.5 5.0 1.250 4.0 18.5 4.613 2.5 18.0 4.500 2.0 19.5 4.845 0.5 7.0 1.750 3.5 18.5 4.613 1.5 14.5 3.613 3.0 17.0 4.250 1.5 17.5 4.375 3.0 24.5 6.125 0.832 0.357 0.322 0.254 0,281 0.388 0.373 0.404 0.418 8.3 4.8 3.5 11.2 3.3 36.7 11.5 10.4 7.7 9.3 12.0 13.8 9.9 17.1 17.4 13.2 8.0 29.4 6.9 76.5 30.0 20.4 19.7 13.3 24.9 29.7 34.2 41.1 354 s. KUSANO : TEANSPIRATION OF TABLE V (January 31-February 4).* Only on the first day, the glass cover was removed. The weather was fine throughout the day. Weighings were made at 4 p.m. on the first and the last days. Names of plants. Cryptomeria ... Pinus Podocarfus ... Torreya Chamcccyparis Quercus Pitlospoj-um ... Illicium Tèmstrœmia ... Thea Eriobotrya Photinia Thlsia DaphniphyUum Total amount of transpiration. 23.0 26.0 31.0 10.0 27.5 34.5 29.5 12.0 32.5 24.5 30.5 14.0 32.5 Average of daily amount of transpiration. 5.750 6.500 7.750 2.500 6.S75 8.625 7.375 3.000 8.125 6.125 7.625 3.500 8.125 Transpiration during 21 hours. per n dm. of surface. 1.238 0.684 0.491 0.436 0.495 0.658 0.66S 0.323 0.555 per 100 gr. of fresh weight. 5.4 5.7 19.8 6.6 54.7 22.1 15.9 13.1 16.3 20.3 21.8 7.9 22.7 per 100 gr. of dry weight. 14.8 13.0 52.0 13.8 113,7 57.6 31.0 33.8 23.5 42.3 53.2 27.4 54.5 1st fMean 2..3 * Tcmperaturo-j Maximum ... 8.4 (Miiiiimnn ...-3.2 Relative huiiiitlity 45.0 2nd 3rd 4th. 2.9 3.9 .3.4 8.1 9.5 9.8 -3.5 -0.6 -2.0 40.5 43.5 .58.0 EVERGREEN TREES IN WINTER. 355 > a ^-l 1-; o P-, -ri '^ o •+J c3 1—1 g- CI C3 4 1—1 p s To is ^ > CO K- e rQ *r^ 0) [bio to f^ ■i^ ^ d n ^ <1 S 2 ci o p: o o ■t-^ Ö Oi r^; CD c3 JO JO JO •rap □ .xad TS pu« 0?; 'ei JO noiiujidsinj.q JO 'junoiau "^n^'P JO sSB.xaAy •UOI)B.IldsUTJ.I} JO -junoraB ^[Tlîp JO 8glJ.iaAY C^OiMff-lr-jOi— lOMOliCOfMO eocöoïoö(^ic■^o5c<^I-^t>^eÖT^^(^5ö CS Cl iH -^ 1— I lO CO CO CO 1— I t-- CO CS «0 t-.iofOt-.osMjajiO'^'^cot^io Cr> CO OÖ lO (>i CD Oi lÖ c4 lO oj cö t--' >— I t^ IM 1— I .— I 1— I CO — I ■— ( [, Ttl O C5 cqcocoioj^i-Hooo l-î Ö Ö Ö Ö "-^ Ö rH fcîî'^*^l^t^t^*>'OI:^t^t--OC0 COCOCOCOCOCOCOCOOCOCOCOOC- T-;cOcoi-;i-HT-;i-Ht-HiOi-Hcqi— looco cooc5(NCi0500Jeocööor-Io5 CO Tfl rH -rtl ,— I ,— I t^ lO 1- t^ to I— I CO !>. •^ O t^ O I— I O i-l >-( o lo o t^ cr> L^ CO iq CO CO o iq lO CO c<5 o5 oö t-^ go' t~^ iq o iq iq o O T-H T^ oi CO i-< CO rfi lO ^ '^i O O lO O O lO lO o Ttl LO -"Jl rH o lO lO o lO LO lO o o lO lO o LO lO o o Kt) 03 > 10 I) )) 11 15. » 12 11 )) 13 16. » 14 » J) 15 17. )) 16 18. )) 17 19. )) 18 21. » 19 22. ,, 20 11. Jan. 21 17. !I )) 18. )) 22 20. )> 23 23. )) 24 22. )J 25 27. » 26 28. » 27 » )) Fine Cloudy Fine Cloudy Fine Rainy Fine Cloudy Fine Cloudy Fine Cloudy Aucuba japonicn'f Pittosporum Tobira Ligularia Kœmpferi Quercus glauca ■ Thea japonica Pascmia cuspidaia , Ilex crenata Photlnia glabra Gardenia florida Daphniphjllum macropodum.. Clnnamomum Loureirii Ligustrum japonicum Ternstroinia japonica Eriobolrya japonica Thea S'ascmqua Nandina domesiica Aspidistra elatior Gymnogramme japonica .... Podocarpus macrophylla .... Fatsia japonica Daphniphylluvi macropodum . Aucuba japonica , Pasania glabra Tliea japonica Ternstnnnia japonica Aucuba japonica Pittosporum Tobira .. 1 G 1 15 1 1 3 19 4 12 1 10 7 172 3 23 4? 35 1 16 1 11 5 60 b 42 2 9 5 29 1 130 1 1 1 21 5 64 1 1 1 13 )) !) 1 10 1 9 5 28 4 48 1 9 2 63 18.812 3.372 8.077 10.640 7.940 7.152 9.550 9.870 19.840 13.500 28.220 9.830 5.055 9.525 8.700 8.930 35.370 10.265 18.010 17.G55 15.410 17.257 9.942 12.759 9.322 Ddiu. 6.794 0.833 1.686 4.007 1.965 2.217 2.251 3.347 5.263 5.862 5.433 8.195 2.876 1.878 2.680 8.197 3.186 5.456 6.166 2.570 5.518 )) 5.510 4.407 4.386 2.028 4.132 3.925 C. 11.5 10.0-10.5 10 8.8 9.2 9.4-9.5 9.9-10 9.9 9.1-9.2 9.6-9.7 8.7 9.0 9.2-9.0 9.3-8.8 10.2 9.2-9.0 7.7-7.8 2.9-9.1 7.4 9.0-8.0 0.6 1.8 2.6 1.4 2.0 2.6-2.8 2.8 3.2 c. 12..5-12.6 11.0 10.5-11.0 10.0 10.0 10.0 11.0 10.2 10.0-10.2 10.5 9.5 10.0 9.2-9.6 10.0 10.0 8.9-9.0 8.0-8.2 9.0 6.4-6.5 8.8-9.0 0.0 1.0 2.0 0.5 1.0 1.5 2.0 2.0 * One millimeter of water column in the capillary tube corresponds t The branches were cut off on the preceding day and placed in t The branch had been cut off 4 hours before. § The absorbing surface had been renewed 4 hours before. y The branches had been cut oil' 3 hours before. EVERGREEN TREES IN WINTER. 359 Branches under Diffused Light. IX. is lining of It. Column of water in the tube absorbed o .2 --^ 2 'S in eacli succeeding interval of 'Is 1 t f tran.s m. pet 10 minutes.* 3 '^ o o r; '^ « i H a 0- ■«1 84.6''" 11.00 a.m. 80, 80, 81, 80, 79, 79. mgr. 371.225 54.64 85.1-S7.7 o.lO p.m. 11, 10, 10, 9.5, 9.5, 9.5. 64.112 55.36 84.3 3.10 „ 26, 26, 26, 26, 26, 26. 120.900 71.71 81.2 11.00 a.m. 84, 89, 89, 81, 79, 79, 78. 447.175 95.66 75.9-78.5 2.20 p.m. 28.5, 29.5, 28, 28, 28, 24, 23, 25, 25, 24. 203.826 62.24 80.6 10.50 a.m. 21, 20.5, 20, 19.5, 18, 19, 20. 10G.950 41.30 84.3-86.8 3.00 p.m. 12.5, 12.5, 12.5, 11, 11, 12, 12.5, 11.5. 74.013 24.62 72.9 11.20 a.m. 25.5, 25, 23, 23, 23, 21. 108.888 32.53 81.1-81.2 10.40 „ 67, 70, 67, 66, 63, 57, 60, 58, 57. 439.375 53.54 78.9-80.4 3 00 p.m. 85, 82, 79, 79, 78, 79. 373.550 63.72 79.4 11.00 a.m. 45, 48, 47, 45, 46, 45, 46. 259.650 40.95 82.4 3.00 p.m. 63.5, 54, 55, 55, 51, 58, 55. 301.413 31.53 79.6 11.00 a.m. 18, 19.5, 18, 18, 19, 20, 20, 19, 19. 146.470 30.56 74.8-79.7 2.30 p.m. 28, 29, 25, 24, 22, 21, 22, 22, 22. 166.6-25 59.15 77.1 2.10 „ 52, 46, 48.5, 45.5, 48, 45, 44. 254.976 81.55 69.8-73.1 2.40 „ 81, 88, 83, 82, 77, 96, 64, 78. 502.975 46.02 84.4 12.20 „ 4.5, 5, 4.5, 4, 5, 4, 4.5, 4. 27.514 6.48 69.4-64.6 2.00 „ 110, 108, 105, 110, 110, 114, 116, 116, 114, 110. 802.575 96.86 69.6 3.00 „ 85, 79, 71, 69, 57, 69, 58. 378.200 52.57 56.6-72.2 1.00 „ 28, 28, 30, 33, 37, 29. 143.375 55.79 90.0 12.30a.m.t 30, 30, 29, 25, 26, 25, 25. 147.250 22.87 67.8 11.30 „§ 25, 25, 25, 25. 77.500 21.07 86.8 11.00 „II 18, 16, 16, 16, 16, 15. 75.175 13.60 62.8 11.20 „II 10, 9, 9, 10, 9, 8. 42.625 9.90 68.7 11.20 „ II 23, 26, 24, 24. 75.175 25.71 91.0-74.6 10.30 „ 12, 12, 11, 10, 10, 10. 50.375 24.84 72.1 10.30 „ 13, 12, 12, 12, 12, 12. 56.575 13.69 72.6 1.10 p.m. 19, 19, 18, 16, 16, 17, 15, 15. 104.625 19.99 to 0.775 mgr. water in the room until the experiment was made with new cutting surface. 360 s. KUSANO : TRANSPIRATION OF III. Mode of the Absorption of Water by Cut-Branches AT Low Temperatüre. TABLE X. January 17, 1899— Weather very fine. Plant : Dcqjhniphyllum macropochim. Air temperature, at the time when the branch was cut ofi", was 0.8*^0. and the leaves were drooping. Time of preparation — 8.30 a.m. Time. Column of water absorbed in mm. Air temperature. "Water temperature. Remarks. 9.00 a.m. 0.2 0 9.10 1 0.2 0 9.20 1 0.2 0 9.30 1 0.2 0 9.40 1 0.2 0 9.50 2 0'.2 0 10.00 2 0.2 0 10.10 3 0.2 0 10.20 3 0.2 0 10..30 4 0.2 0 10.40 73 0.6 0 Inclination of a leaf | 10.50 223 0.6 0 was 65.° 11.00 245 0.6 0 11.10 180 0.0 0 11.20 110 0.6 0 11.30 80 0.6 0 11.40 57 0.6 0 11.50 58 0.6 0 70.° 12.00 44 0.6 0 12.10 p.m. 38 0.6 0 78.° 12.20 35 0.6 0 12..S0 31 0.6 0 12.40 30 06 0 81.° 12.50 30 0.6 0 1.00 29 0.6 0 1.10 25 0.6 0 1.20 26 0.6 0 1.30 25 0.6 0 1.40 25 0.6 0 EVERGREEN TREES IN WINTER. 361 TABLE XI. January 18, 1899 — Weather very fine. Plant : The same branch after the first experiment was placed in water outside the laboratory during the night until tlie next morning and then after making a new absorbing surface experiment was repeated. Air temperature at 7 a.m. was 1.1°C. and temperature of water in which the branch remained immersed was 0°C. Time of preparation — 7.30 a.m. Time. Column of water absorbed in mm. Air temperature. Water temperature. Remarks. 8.50 a.m. 5 1.2 0.3 At the beginning of ob- 9.00 4 1.2 0.3 servation, i.e., at 7.30 9.10 3 1.2 0.3 a.m. the temp, of air 9.20 4 1.2 0.3 and water in the 9.30 4 1.2 0.5 room was 1°C. and 9.40 3 1.2 0.5 0°C. respectively. 9..50 7 1.2 0.5 Inclination of leaves 10.00 25 1.4 0.5 remained constant 10.10 115 1.4 0.5 during observation. 10.20 82 1.4 0.5 10.30 Gi 1.4 0.5 10.40 53 1.4 0.5 10.50 45 1.4 0.5 11.00 30 1.6 0.5 11.10 33 1.6 0.7 11.20 32 1.6 0.7 11.30 25 1.6 0.7 11.40 25 1.0 1.0 11.50 25 1.6 1.0 12.00 25 1.6 1.0 362 s. KUSANO : TRANSPIRATION OF TABLE XII. January 20, 1899— Weather very fine. Plant : Aucuha Japon ica. Air temperature, O'^C. at 8 a.m. The leaves drooped and were curled up. Ten minutes after they were brought in the room, all the leaves became turgescent. Time of preparation— 8.05 a.m. Time. Column of water absorbed in mm. Air temperature. Water temperature. Remarks. 8.30 a.m. — 2.4 2 A leaf took horizontal 8.40 5 2.4 2 position. 8.50 G 2.4 2 9.00 5 2.4 2 9.10 G 2.4 2 9.20 11 2.5 2 9.30 12 2.5 2 Here it took the normal 9.40 26 2.G 2 erect position. 9.50 50 2.6 2 10.00 35 2.6 2 10.10 29 2.6 2 10.20 23 2.6 2 10.30 22 2.6 2 10.40 20 2.6 2 10.50 14 2.6 2 11.00 23 2.6 2 11.10 18 2.6 2 11.20 IG 2.6 2 11. .30 IG 2.6 2 11.40 16 2.6 2 11.50 16 2.6 2 12.00 15 2.6 2 EVERGREEN TREES IN WINTER. 3G^ TABLE XIII. January 28, 1899— Weather cloudy. Plant : Aucuba japonica. The branch was brought from open air at -4°C. into the room at 2.6°C. The leaves were covered with frost. Time of preparation— 8.30 a.m. Time. Column of water absorbed in mm. Air temperature. Water temperature. Remarks. 8.30 a.m. — 2.7 2 8.40 2 2.7 2 8.50 1.5 2.7 2 9.00 1.5 2.6 2 9.10 3.5 2.5 2 9.20 6.5 2.5 2 9.30 6 2.4 2 9.40 14 2.4 2 9.50 20 2.4 2 10.00 18 2.5 2 10.10 14 2.5 2 10.20 14 2.5 2 10.30 13 2.7 2 10.40 12 2.7 2 10.50 12 2.8 2 11.00 12 2.8 2 11.10 12 2.8 2 11.20 12 2.8 2 364 s. KUSANO : TRANSPIRATION OF TABLE XIV. January 27, 1899— Weather very fine. Plant : Tcrnstrœmia japo'iKica. The branch was brought from open air at 0°C. into the room at 2.2°C. Time of preparation— 8.20 a.m. Time. Column of water absorbed in mm. Air temperature. Air temperature. Remarks. 8.30 a.m. — 2.3 1.0 8.40 12 2.4 1.0 8.50 72 2.4 1.0 9.00 57 2.4 1.5 9.10 44 2.4 1.5 9.20 31 2.4 1.5 9.30 25 2.4 1.5 9.40 21 2.4 1.5 9.50 20 2.5 1.5 10.00 17 2.5 1.5 10.10 IC 2.5 1.5 10.20 15 2.0 1.5 10.30 13 2.6 1.5 10.40 12 2.0 1.5 10.50 12 2.6 1.5 11.00 11 2.6 1.5 11.10 10 2.8 1.5 11.20 10 2.8 1.5 11.30 10 2.8 1.5 EVERGKEEN TEEES IN WINTER. 365 TABLE XV. January 23, 1899 — Weatlier very fine. Plant : Fasania (/labra. Air temperature out uï doors was -S°C, at 6 a.m. month. Time of preparation — 7.20 a.m. the minimum of this Time. Coliiinn of water absorbed ill nun. Air temperature. Water temperature. Remarks. 7.30 a.m. — — _ 7.40 — — — 7.50 16 0.4 0.5 8.00 IG 0.4 0.5 8.10 16 0.4 0.5 8.20 21 0.4 0.5 8.30 23 0.7 0.5 8.40 25 0.8 0.5 8.50 20 1.0 0.5 9.00 17 1.0 0.5 9.10 14 1.0 0.5 9.20 12 1.0 0.5 9.30 12 1.0 0.5 9.40 11 1.0 0.5 9.50 10 1.0 0.5 10.00 12 1.2 0.5 10.10 10 1.3 0.5 10.20 9 1.3 0.5 10.30 10 1.3 0.5 10.40 9 1.1 0.5 10.50 9 1.3 0.5 11.00 10 1.4 0.5 11.10 9 1.4 0.5 \ßß Contents. I. Introductory. II. Metliod. III. ïlie Climate of Micldle Japan. IV. Evergreen Trees of Japan. V. Transpiration under Direct Insolation. VI. Transpiration under Diffused Light. VII. Summary. Experimental Data. I. Transpiration by Direct Insolation of Pot-Plants. II. Transpiration of Cut-Branches in Diffused Light. III. Mode of the Absorption of Water by Cut-Branches at Low Tem]i.crature. '^ \ b; ^ ? §• a' ^• Ci ^ «■ g - ■ "^^ p ^ jg 5 Ç^ t Î s 5 0 3 0 i .3 Ueber die Sporocarpenevacuation und darauf erfolgendes Sporenausstreuen bei einer Flechte. VON M. Miyoshi, Higcihuhaknshi, Professor der Botanik a. d. Kaiserl. Univers. z. Tokio. Mit Tafel XVIII Bis. Bekanntlicli entleeren die gymnocarpisclien Flechten ihre Sporen leicht, wenn die Oberfläche der Apothecien befeuchtet wird, so dass das stark gequollene, nach aussen gewöll^te Hymenium durch den Bruch der Ascuswandung die Sporen mit GcAvalt ausschleudern lässt. Anders verhält es sich mit den angiocarpischen Flechten, deren Hymenium nur durch ein kleines Loch nach aussen geöffiiet ist. Hier verhindert das harte Perithecium eine starke Volumen- zunahme des gequollenen Hymeniums, vermöge dessen Druck die Sporenmasse aus dem Ostiole ausgetrieben wird. Bei denjenigen Angiocarpen, welche mit keiner natmlichen Öjffnung versehen sind, muss die Sporenentleerung nur durch den Bruch des Peritheciums an einer Stelle der Aussenwand statt- finden ; aber es ist meines Wissens der Fall nicht genügend be- kannt, dass der ganze Sporocarp mit oder ohne Stückchen des umgebenden Peritheciums von dem Thallus sich lostrennt, abfällt und zur Sporenentleerung bereitet wird. 368 M. MIYOSHI : UEBER DIE SPOKOCARPENEVACUATION UND Im Januar 1898, gelegentlich einer botanischen Excursion nach Idsu, fand ich eine Krustenflechte auf der Kinde von Elceocarpus decipiens Hems. Nach Untersuchung erkannte ich die Flechte als eine neue und nannte sie Sagedia macrospora mit folgender Diagnose : Kruste verbreitert, dünn, häutig-schorfig, graugrün. Früchte fast von der Kruste bedeckt, halbkugelig, bis 1 mm gross, schwarz. Hymenium farblos, einfach, von schwarzem Perithecium um- schlossen. Schläuche cylindrisch, 8-sporig ; Sporen spindelförmig, bisweilen gekrümmt, beide Enden scharf gespitzt, vieltheilig, ca. 170 /i lang, 18/^ breit. Gonidien gelbgrün. Unsere Flechte zeichnet sich aus : erstens durch ihren mattgrünen, sehr verbreiterten Thallus, welcher oft eine Kreisfläche von mehr als 1 Meter in Durchmesser bedeckt, zweitens durch die Art und Weise ihrer Sporenentleerung, welche ich hier aus- führlich beschreiben will. Beschaut man die Thallusoberfläche, so findet man neben zahlreichen, durch Thalluskörper beinahe bedeckten, intakten Sporocarpen hie und da glatte, weisse Höhlchen, welche nichts anderes sind als die früheren Stellen der vom Thallus bereits losgetrennten, abgefallenen Sporocarpen. Davon, dass der Fruchtkörper leicht vom Thallus abgesondert werden kann, überzeugt man sich gleich, wenn man mit einer Nadelspitze ihn herauszunehmen versucht. Dann trennt sich die ganze Masse des Sporocarps glatt von dem umgebenden Thallustheile mit oder ohne Begleitung eines Theilstückes des letzteren. Besonders leicht gelingt diese Operation bei den völlig ausgebildeten Sporocarpen, welche an den natürlichen Standorten der Flechte fortwährend vom Thallus getreinit werden und abfallen. Der auf die eben beschriebenem Weise leicht ]3efreite Fruclitkör])er DAEAITF ERFOLGENDES SPORENAUSSTREUEN RET EINER FLECHTE. 360 (Taf. XVIII Bis, Fig. 1) ist von einer weissgelblichen Farbe, waclisartiger Consistenz und kugeliger Gestalt, ea. 72 nun im Durchmesser und Y,o mg im Gewicht. Seine ganze Masse besteht aus einem massiven Hymenium, welches äusserlich grösstentheils, aber besonders am unteren Theile von einer weissgelblichen Hypothecium-Hülle bedeckt ist. Befeuchtet man einen solchen isolierten Sporocarp auf einem Objectglas, so verquillt nach etwa 5 Minuten das FTymenium durch lebhafte Wasserabsorption so sehr und wölbt sich so stark nach aussen, dass die weniger quellbare äussere Hülle zerbrochen und ihre Theilstücke, welche noch fest mit dem Hymenium verbunden sind, von dem letzteren fast ganz nach innen getrieben werden (Fig. 2, 3, 4). Auffallend ist nun zu sehen, wie die langen, feinen Paraphysen sich erstrecken, welche strahlig von einem Centrum sich ausbreitend etwa einem um eine Fliege herumgewachsenen Sap.rolepilanxsen ähnelt. Wenn die Paraphysen sich ausbreiten, so schleudern sich die grossen spindelförmigen Sporen aus den Schläuchen aus, und schreiten mit einer ziemlichen Geschwindio'keit in dem umo'ebenden Wasser fort. Jetzt wird der ursprüngliche Durchmesser des Hymeniums bedeutend vergrössert, etwa bis V/. mm, so dass das Volumen sich ums 27 fache vermehrt. Unter Wasserausziehung entweder durch einßiches Trocknen- lassen oder mittelst Alcohol, resp. Glycerin, contrahiert sich das Hymenium stark nach innen und lässt die äussere Hülle ihre frühere Stelle wieder einnehmen. Dank der schleimio-en Eisen- Schaft des Paraphysen wird das Hymenium an der Unterlage fest angekittet und somit seine Haftstelle gesichert. Das Aufquellen und Zusammenziehen durch Auf- resp. Ab- nahme des AVassers kann wiederholt veranlasst werden und 370 M. MIYOSHI r UEEER DIE SPOROCARPENEVACUATION &C. sogar bei solchen Sporocarpen, die mittelst Alcohol oder Hitze )3ereits getötet worden sind. Die Quellungskraft des Hymeniums muss bedeutend gross sein. Sporocarpen, die in 5*^0 iges, im Erstarren begrifTenes Gelatine schnell hineingelegt waren, quollen endlich füst zu normaler Grösse ; in lO^^igem Gelatine fand die Quellung nur langsam statt, und in 20'^oigem und höher-procentigem Gelatine geschah eine Zeit lang nach der Einbettung fest keine Quellung mehr.^) Aus den oben beschriebenen Thatsachen ersieht man leicht die biologische Bedeutung des eigen thümlichen Verhaltens des Sporocarpes. Vermöge einer leichten Trennbarkeit vom Thallus fallen die winzigen Fruehtkörper in der Umgebung ab und werden nun durch Wind auf andere Baumrinden fortgeführt. Werden sie durch Kegen oder Thnu benetzt, so quellen sie sofort auf und streuen ihre Sporen aus. Die Klebrigkeit des schleimigen Hyme- niums hilft dem letzteren, sich leicht an den Baumrinden anzuheften. Es muss aber bemerkt werden, dass bei unserer Flechte die gewöhnliche Sporenentleerungsweise auch stattfindet, indem der schwarze Scheiteltheil des Sporocarpes nach der Keife einen Bruch erfährt und eine offene Mündung nach aussen bildet, wodurch die Sporenmasse nuf die übliche Weise ausgestreut werden kann. 1) Eine gcimiipre Messung der Qiiellungskrafl imite ieii niolit gemnclit. Ueher die Unter- sncliungsmetliodik der Qiiellung der PHanzenlst'irper vergl. man Rktnke's lielcannte Aiiliandlnng in IIan^tein's Hotani^clien Aliliandlinigen lîd. IV, lieft I, 1S79. TAFEL XVIII B.s. Tafel XVIII lîis zeigt n;icli eiiKinder folgende Stadien der (iuellung eines Sporocarps von Sd'jcdia macroapora. Die Plguren wurden unter meiner Aufsicht von Herrn I. N is hi no mittelst Camera lucida gezeichnet. Vergrosserung ca. 35 mal. Fig. 1. Ein Sporocarp vor der Quellung. Fig. 2. Derselbe in einem anfänglichen Stadium der Quellung. Der obere Tlieil des Hymeniums ist im Begriöe durch Wandbruch auszutreten; Paraphysen, Sjjoren und Luft- blasen sind zu sehn, Fig. .3. Ein weiter fortgeschrittenes Stadium, indem eine grössere Masse des Hymeniums durch starke Voluraenzunahme aus der Wandhülle herausgetreten ist. Fig. 4. Ein stark gequollener Sporocarp wie in Fig. 3, aber von hinten gesehen, mit zahlreichen Sporen auf dem weissen Hymeniumgrund und der auswärts gekrümmten Wand- hülle. Jour. Sc. Coll. Vol. XV., PI. XVIII Bis. Figs. Fij^.4. Studien ueber die Einwirkung des Kupfersulfats auf einige Pflanzen. VON H. Hattori, Rigakushi. Mit Tafel XIX. I. Einleitung und Litteratur. Bekanntlich sind die Einwirkungen der Kupfersalze auf den Pflanzcnköiper je nach den Organen und Entwicklungsstadien weit verschieden : so können z. B. die Samen ihre Keimfähigkeit nach EinÜuss ziemlich konzentrierter Kupferlösungen noch l^ei- behalten, dagegen sind die Keimpflanzen, insbesondere ihre Wurzel- chen mehr empfindlich und werden leicht beschädigt. Oefters findet sich aber Vegetation da, wo Kupfer in einer beträchtlichen Menge in der Erde vorkonnnt und doch bleiben die Pflanzen sanimt ihrem Wurzelsysteme relativ unbeschädigt, da die Bodenerde dank ihrem grossen Absorptionsvermögen für Metallsalze als ein kräftiges entgiftendes Mittel dient.') Zudem zeigen die Ver- suche von Viala,") welcher eine Topferde drei Monate lang 1) Pfeffer, Pflanzenphysiologio, Bd. I, AuH. 11, 1898, \k 148 und 421). 2) Viala, De l'actiuii de ceituiue« subblances toxiques su r la vigne. Kef. J u^ I, Jalires- Leiicht, 1895. 372 H. HATTOEI : EINWIRKUNG DES mit einer Kupfervitriollösung begossen, und die dadurch im Boden incorporierte Menge des Kupfersulfats hoch ansteigen Hess, dass trotz dieses grossen Gehalts an Kupfer, die Eebe gesund blieb. Anderseits liegen Angaben vor, dass Kupfer in einer grösseren oder o-erinfiferen Men ne ohne sichtbaren Schaden im lebenden normalen Pilanzenkörper vorkommt.^) So fand Lehmann") dass die in der Nähe ein Kupferwerkes erwachsenen Pflanzen eine nicht unwesentliche Quantität des Kupfers (83 bis 560 mg. in 1 kg Trockensubstanz) ohne besondere Beeinträchtigung des Lebenspro- cesses aufnehmen können. Tschirch"^) sagt in seinem bekannten Werk über Kupfer wörtlich dass ,, die lebende Pflanze Kupfer- sowohl durch die Wurzeln als auch durch die Epidermis aufzunehmen im Stande ist und auch immer aufnehmen wird, wenn es ihr im Boden dargeboten wird,'' und er fand dass ein auf kupferhaltigen Boden erwachsener Weizen in einem Falle in Ü50 gramm der Ernte 0.2775 CuoS enthielt. Solche Thatsachen sind ferner durch die einschlägigen Versuche von Phillips/) Freytag,^) B er le se und S OS teg ni,^) bekannt geworden, nämlich, dass Kupfer in Boden mehr oder minder von der Pflanzen ohne Schaden absorbiert werden kann. Betreifs der Kupfervergiftungsversuche bei Wasserkulturmethode liefert die Arbeit Hasel h off' s^) anderseits 1) rfcfier, I.e. p. 432. Vergl. auch Mac. Doiigal, Eut. Gaz. Vol. XXVII, 1899, p. 68-69. 2)Lelimann, Hygienische yiiulien über Kupfer, IV. Archiv f. Hyg., Bd. XXVII, 189t). o)TrtchircJi, Das Kupfer. 1893, ]). 15-17. 4)Pliillil)s, On the Absorption of Metallic Oxides f)y Plants. Cliem. News, 1882. 5) Freytag, Die schädlichen Bestandthcile der Hüttenrauchs der Cu-, I'b-, Zn-IIütlen und ihre Beseitigung. Landw. Jalirb. Bd. XI, 1882. 6) Berlese et Sostegni, Kecerclies sur l'action des sels de cuivre sur la végétation de la vigne et sur le sol. 1895, Sond. Al)d. aus La Revue international d. viticulture et d. Oenologie, 1895. 7)lIaseIhuH, l'elnn- dio sch;idli< lie Wirkung des Kupfersulfat und Kupfernitrat halligen Wassers auf Boden und l'ljanzen. Landw. Jahrb. Bd. XXI, 1893. KUPFERSULFATS AUF EINIGE PFLANZEN. 1373 einen Beweis dafür, dass beim Mais die scliädliclie Wirkung des Kuptersulfîxts bereits bei 5 mg CuO pro 1 Liter Nährlösung beginnt, bei Bohnen hingegen erst bei 10 mg. Otto^) gelangte ebenfalls zu ähnlichen Ergebnissen. , Im Jahre 1893 erschien die interessante Arbeit Naegeli's^) über oligodynamische Erscheinungen welche zeigt, dass eine äusserst geringe Menge Kupfers eigenthümliche Absterbenserschei- nung an Sini^ogyra hervorbringt. Dieselbe Thatsache wurde von Cramer) und Bumm'*) weiter geprüft. In neuester Zeit stellten Kahlenberg und True') und gleichzeitig HeakF') eine Beihe von Versuchen mit sehi" verdünnten Metallsalzen und organischen sowie anorganischen Säuren bei Lupinus, Zea, Pisum und Cucurhita an und gelangten zu dem Ergebniss, dass die Ursache der Giftwirkung solcher verdünnten Lösungen auf die dissocierten Ionen derselben meistentheils auf den Cathionen beruht. Was die Wachsthum beschleunigende Einwirkung einiger Substanzen durch chemische Beize anbetrifft, so liegen uns LTnter- suchungen von Bfeffer^) vor, welche beweissen, dass die Lebens- vorgänge durch kleine Mengen gewisser besonders auch giftiger Stoffe beschleunigt werden können. Die Behandlung mit Kupfer- l)Otto, Untersuchungen über das Verhalten der Pflanzen-Wurzeln gegen Kupfersnl- fatlüsungen. Zeit. f. Pfl. Krankh., Bd. III, 1893. 2) V. Naegeli, Oligodynamische Erscheinungen in lebenden Zellen. J893. 3) Cramer, Nachtrag zu Naegelis Arbeit. 4)Kuram, Zur Kenntniss der Giftwiikung der Bordeaux-brühe &c. Zeit. f. wiss. Bot., Ed. T, 1895, p. 99. 5) Kahlenberg and True, On the toxic Action of dissolved Salts and their electric Dissociation. Bot. Gaz., Vol. XXII, 1S9G und auch Copeland and Kahlenberg, The Influence of the Presence of pure Metals upon Plants. Trans, of Wisconsin Acad , Vol. XII, 1899. 6) Heal d, On the toxic Effect of dilute Solutions of Acids and Salts upon Plants. Bot. Gaz., Vol. XII, 1896. 7) Pfefler, Ueber Election der oiganifchen Nährstoffe. Pringsh. .Jahrb. f. wiss. Bot., Bd. XXVIII, 1895. 374 H. HATTOEI : EINWIRKUNG DES kalkmischung, der sogenannten Borcleauxbrühe, z. B. liât eine günstige Wirkung woraus schon einige Forscher besonders Eumm/) Frank und Krüger") und AderhoUF) aufmerksam gemacht haben. Diese Erscheinung schrieb Rumm einem chemi- schen Reiz zu. Es ist aber noch nicht klar gestellt, welcher Bestandtheile der Brühe solche beo;ünsti2:ende Einwirkuno- auf Pflauzen ausüben kann. Von neueren Untersuchungen über die Reizwirkung von Metallsalzen auf Schimmelpilze sind besonders die von Richards'*) und von Ono') zu erwähnen. Letzterer hat zuerst auch für Kupfer eine solche Reizwirkung dargethan. Interessant und wichtig ist nun festzustellen in welchen Ver- hältnissen das Kupfer als Gift resp. als Reizstoff wirkt und wie seine Wirkung durch obwaltende Bedingungen beeinflusst wird. Desshall) habe ich die vorliegenden Untersuchungen von September 1898 bis zum Juni 1890 in botanischen Institut der Kaiserlichen Universität zu Tokyo, unter Leitung des Herrn Prof. Dr. Miyoshi angestellt, und kam so weit meine Versuche erlaubten zum Ergebnisse, dass die Kupfervitriollösung in weit verdünnterer Konzentration z. B. 0.00005-0.0000059^, als die von Haselhoff, Otto und Anderen verwendete, eine auffallendes Gift für Wurzeln einiger höherer Pflanzen ist, und fernei', dass die Giftwirkung wie 1) Rum in, Ucbcr die Wirkung der Kupferpreparate bei Bekämpfung der sogenannten r.lattfallkranklieit der Weinreben. B. d. I). B. G., Bd. XF, ISO,*?. Kumm, Zur Frage nacli der Wirkung der Kupferkalk-Salze bei Bekämpfung der Peronospora viticola. B. d. D. B. G. B.d. Xf, lS9;î. 2) Frank und Krüger, Ue})er die Reize welchen die Bakterien &c. B. d. D. B. G., B.d. XII, 1894. o) A der h old, Ueber die Wirkungsweise der sogenannten Bordeaux-brühe. Centralblatt f. Bakt. &c. II. Abt. 15d. V, 1899. 4)H. M. Richards, Die Beeinflussung des Waehsthums einiger Pilze durch cheniisclie Eeize. Pringsh. Jahrb. f. wiss. Bot., Bd. XXX, 1897. 5)Ono, Ueber die Wachsthumbeschleunigung einiger Algen und Pilze durch chemische Keize. Jour. Coll. Sei., Imp. Univ., Tokyo, Vol. XIII, 1900, p. 141. KUPFERSÜLFATS AUF EINIGE PFLANZEN. 375 bekannt durch das Absorptionsvermögen des Bodens mehr oder weniger vermindert werden kann. Betreffs der wachsthumbesch- leunigenden Einwirkung von Kupfer auf einige Schimmelpilze steht mein Versuchsergebniss mit demjenigen von Ono im grossen Ganzen im Einklang, iudem das Trockengewicht der mit geeigneter Dosis (z. B. bei Aspergillus niger 0.004^ ) des Kupfers kultivierten Pilzes demjenigen der Konti'ollkultur gegenüber bis zum Doppelten gesteigert wurde. II. Methodisches. In uuseren Versuchen mit Ausnahmen von denjenigen bei Topfpflanzen, in welcher das gewöhnliche chemisch reine Kupfer- salz verwendet war, kam Merk's garantiert reines Kupfersulfat (CuSO^ + öHoO) zur Anwendung. Bei Versuchen mit Zweigen wandte ich Glassgefösse von ca. 2 Liter Inhalt mit gut seh liessenden, in der Mitte durchbohrten, Korkstöpseln an. Die Versuchsobjekte von möglichst gleich- massiger Grösse und Aussehen wurden unter Wasser abgeschnitten, und durch das Loch des Korkes in die Lösung gesteckt. Die Versuche mit Topfjiflanzen führte ich mit solchen Exemplaren aus, welche mehrere Monate lang in Töpfen gepflanzt gewesen waren und in Grösse und Gesüdt von einander nicht wesentlich abwichen. Die Topfpflanzen befanden sich im Freien und wurden jedem Tag mit einer bestimmten Menge Cu-Lösung begossen. Behufs Bestimmung der minimalen Konzentration, bei welcher die erste sichtbare Schädlichkeit zu beobachten ist, stellte ich die Versuche in AVasserkulturen in ca. 2 Liter, haltigen Glassgefässen 376 H. HATTOKI : EINWIRKUNG DES an. Die Samen wurden nachdem sie zum Aufquellen etwa 24 Stunden lang unter destilliertem Wasser verweilt und in Säge- spänen zum Keimung gebracht waren, bis die Wurzeln etwa 1-2 cm lang geworden waren, auf über Wasser gespannten Fadennetzen zur weiteren Entwicklung geln'acht. Nachdem die Keimlinge einige Centimeter erreicht hatten wurden die Pflänzchen aus dem Keimbette entfernt, und in AVasserkultur gezogen. Die Reinigung der Kulturgefasse geschah zuerst mit Salzsäure und dann mehrmals mit destilliertem Wasser. Als Kulturflüssigkeit diente mir nur aus Glas destilliertes Was- ser mit Zusatz von l^estimmten Mengen des Kupfersulfats. Dass ich in allen Fällen stets das reine Wasser, nicht aber Nährlösung anwendete, hat seinem Grund darin, dass das Phosphat welches in einer Nährung unentbehrlich ist, mit ICupfer unlösliches Kupfer- phosphat liildet, und je nach]^dem relativen Mengenverhältnisse entweder den Nährwerth des Phosphors vernichtelt oder die Gift- wirkung des Kupfers aufgehoben wird. Auf diese Weise konnten die Versuche seilest verständlich nur so lange andauern als die Selbsternährung der Keimlinge aus ihren eigenen Reservestoff- behältern ausreicht. Ich stellto die Lösungen von Kupfersalz dadurch her, dass ich eine Originallösung von O.OlYo Gehalt auf den gewünschten Grad verdünnte. Mit der Versuchsflüssigkeit füllte ich Glas- cylinder und l)efestigte eine Keimpflanze durch die Oeffnung des Korkes unter Zuhilfsnahme von Watte derartig, dass die Wurzel in die Lösung eintauchte, während das Endosperm oder Cotyledon nicht direkt benetzt wurde. Um das Licht von der Wurzel abzuhalten wurden alle Gefiisse mit schwarzem Papier um- gewickelt. Um zu beurtheilen ob die Wurzelzellen todt oder noch lebend KUPFERSULFATS AUF EINIGE PFLANZEN. 377 waren, diente mir stets die Plasmolyse-Metliode mit 5^/o Salpeter- lösung. Bei Pilzkultur wandte ich meistens Richards A Lösung^) sowie in einigen Fällen die Pfeffer sehe Nährlösung^) mit Zusatz von 5^/0 Kohrzucker an. Je 100 ccm dieses Gemisches wurde in Erlenmeyersche Kolben von etwa 250 ccm Inhalt gefüllt und üblicherweise sterilisiert. Um auf Kupfer in den Pflanzen zu prüfen, wandte ich Ferro- cyankalium mit schwacher Salzsäure an. Sabatier") hat eine konzentrierte Lösung von Bromwasserstoffsäure zum Nachweis des Kupfers empfohlen, welche selbst mit sehr geringen Mengen Kupfer eine hochrote Färbung liefert. Mehrere Versuche jedoch haben mich veranlasst, dieses Reagens wieder aufzugeben. Es zeigte sich dass das Holz von Coniferen auch bei Abwesenheit von Kupfer damit eine röthliche Färbung liefert. Die Grenze der Empfindlich- keit des ersteren Reagens liegt bei ca. 1/J 00000, in welcher Ver- dünnung des Kupfervitriols das Reagens eine schwach aber doch erkennbare Färbung nach einiger Zeit giebt. Die praktische Empfindlichkeitgrenze des letztgenannten Reagens scheint mir derjenigen von Ferrocyankalium etwa gleich zu sein. III. Das Verhalten abgeschnittener Zweige gegen Kupfersulfatlösungen. In diesem Kapitel will ich die Symptome der an Kupferver- giftung erkrankten Zweige einiger Nadelhölzer beschreiben. Die 1) Richards, I.e. K^HPO, 0.50g. MgSO, 0.25g. NH,N03 LOOg. Eisen Spuren Rohrzucker ...5.00g. Wasser 99.00 ccm. 2)Behren's Tabellen für njikroskopische Technik. III Aufl. 1898, p. 145. Ca(N03), 4 g. MgSO, 1 g. KNO3 1 g. KH„(PO J 1 g. 3) Sabatier, Cumpt rendus de l'académie de science, T. CXVIII, 1893, p. 980 und 1260. 378 H. HATTOKI : EINWIRKUNG DES von mir angewandten Pflanzen waren Pinus Tkunbergil, Crypto- merla japonica und 21iuja japonica, deren Zweige, wie oben er- wähnt, unter Wasser abgeschnitten und dann in die Lösung gestellt wurden. Gewöhnlich ist die eintretende Erkrankung schon nach einigen Tagen zu beobachten, jedoch hängt dies natürlich von dem Konzentrationsgrade und äusseren Bedingungen ab. Untersucht man ein solches Zweigstück clien so zeigt die untere Partie des Siebtheils des Gefässbündels eine bräunliche Färbung, welche von unten nach oben allmählich fortschreitet und endlich bis zum Siebtheil der Nadeln reicht. Zugleich zeigen die Chlorophyllkörper eine Schädigung indem sie ihre normale Gestalt ändern. Demgemäss erhält der Inhalt der Mesophyllzellen eine schwache grüne Färbung und dann degeneriert das Proto- plasma zu einer dunkelbräunlichen JMasse, welche die Bräunung in den Nadeln verursacht. Auf diese Weise beginnt das dunkle, bräunliche Aussehen der Nadeln von der Basis zum Scheitel fortzuschreiten und schlieslich verbreitet sich die Verfärbung auf die ganzen Zweige, und damit geht die Pflanze allmählich zu Grunde. Die Nadeln, besonders diejenigen von Pinus, vetrock- neten unter Bräunnung öfters nur an ihrer unteren Hälfte, wäh- rend die Oberhälfte noch schwach grün blieb. Je jünger die Sprosse desto länger wiederstehen sie dem schädlichen Einfluss. Bei der mikroskopischen Prüfung ergiebt ich, dass die Sklerenchymzellen, die Zellwandungcn tier Basttheile, ferner alle Elemente der Holzkörper bis auf das Mark, je nach der Konzentration der Lösung, mit Ferrocyankalium mehr oder weniger rothbraun gefärl)t werden, aber mit îdjnehmender In- tensität gegen den Gi])fel hin. In dem Holztheil stark mit dem Kupfersalze indjil)ierter Aeste, erfüllen oft die Ferrocyankupfer- Niederschläge die Cu'la.sse und in einigen Fällen, sammelten sie KUPFEKSULFATS AUF EINIGE PFLANZEN. 379 sich in dem Hofräumen der Tüpfel an, älmlicli wie das Eosin in den Versuchen Strasburger's^) über den Verlauf der Leitungsbalinen sich ansammelte. Die in verdünnte Lösung gestellten Zweige zeigen nur in den JMarkstrahlen rothe Färbung, während die Holz- tlieile sehr schwach oder fast vollständig ungefärbt bleiben. Auffallend ist die Erscheinung wie Strasburger") seiner Zeit constatierte, dass die Harzgänge von Finus die in einer Kupfer- vitriollösung z. B. 0.1 Yo oder oft in 0.05^ eingestellt war, im ganzen Holzkörper bis zum Gipfel stets eine schöne grüne Farbe zeigen. Die Intensität dieser Färbung nimmt von der Basis nach den Gi^rfel ab und selbst bei der 0.001 9^ Lösung ist die Farben- reaktion, obgleich schwächer doch immer deutlich bei denjenigen Harzgängen die an den Schnittflächen sich befinden wahrzunehmen. Innerhalb der Nadeln, ist die Farbenreaktion des Kupfers im Hypoderm und in den Gefässtheilen, oft in den Wandungen der Mesophyllgewebe wahrnehmbar. Unsere Beobachtung an Plnus- Nadeln stimmte hierin mit Strasburger 's") analogen Befunden ganz überein. Das Mesophyllgewebe zeigte keine Kupferreaktion bei gewöhnlicher Behandlung mit Ferrocyankalium. Bei diesem Falle sollen wie Loew^) und Tschirch^) betonten, einige compli- cirte organische Verbindungen des Kupfers mit dem Chlorophyll, Lecithin und Fett der Chloroplasten gebildet w^erden, was die Erkennung der Reaction erschwert. Die Zweige, welche in ver- dünnten Lösungen vollständig gesund blieben, wurden verascht und nach Lehmann 's Methode"^) auf Kupfer geprüft, wobei ich stets positive Kesultate erhielt. 1) Strasburger, Bau und Verrichtung der Leitungsbahnen, 1891, p. 579. 2) ebenda, p. 619 u. 622. 3) Strasburger, I.e. p. 6û5. 4)Loe\v, Natürliche System der Giftwirkung, 1893, p. 31 und 34. 5)Tscliircli, Das Kui)fer, 1893, p. 25, 33 u. 35. 6) Le li 111 :i n u , IIygieiii;-che Studiuu über Kuj'ler, IV. Aichiv. f. Hygiene. Ld. XXVIJ, 1896. 380 H. HATTOEI : EINWIRKUNG DES Aus den erhaltenen Resultaten ist zu ersehen, class die mini- male Konzentration der Kupfer vitriollösun gen welche auf die Zweige der 3 Nadelholzarten noch einwirkt zwischen 0.005-0.0019^ liegt. Die Zweige von TJmja sind etwas widerstandfähiger gegen das Kupfer als die zwei anderen Arten, denn erstere konnten bei längerer Zeitdauer in einer O.OOo?^ Lösung fast vollständig gesund bleiben, während die beiden letzteren frühzeitig abstarben oder halb vertrockneten. IV. Das Verhalten von Topfpflanzen gegen Kupfervitriollösungen. Bekannlich kann der Boden eine erhel)liche entgiftende Ein- wirkung auf die Pflanzen bei Salzen schwerer Metalle ausüben. Hierüber habe ich eine Reihe von Versuchen mit Topfpflanzen angestellt und gelangte zu folgenden Resultaten. Jede Topfpflanze wurde, je nach den vorhandenen Erdmengen des Topfes, mit bestinnnten Quantitäten der Kupfervitriollösung und nachher mit reinem Wasser begossen und für eine Zeitlang stellen gelassen. Während der Versuchszeit wurden die Symptome der Erkrankung beobachtet. Aus den von mir ausgeführten Versuchen ersehen wir, dass die angewandten Topfpflanzen {Plnus Thu7ibergii, Thuja occi- denialis, Cryptomerla japonica^), ihre Lebensthätigkeit in solcher Erde, welche mit ziemlich starken Kupferlösungen begossen wurde, auf längere Zeitdauer behalten können, während dieselbe Konzentrationen bei Zweigen, wie unsere vorher mitgetheilten Experimente zeigen, offenljar sehr schädlich gewirkt hätten. Dieser l)I)ic Criii>li)iiiaia-\)[\MVK\\ litten leider clureh einen Sluiiii .so .stark, dass die Resultate der Veröuche.s nicht weiter Leaeiitet werden Ivonnten- KUPFERSULFATS AUF EINIGE PFLANZEN. 381 Unterschied bernlit auf der Verschiedenheit des Mediums, indem die Erde wie schon gesagt wegen ihrer grossen Absorptionsßihig- keit die Giftwirkung der angewandten Lösung erheblich vermin- derte. Ich halje die Al)sorptionskraft der in den Töpfen enthal- tenen Erde für Kupfer durch den folgenden Versuch festgestellt : — 500 gr. luft trockene Gartenerde wurden in 1000 ccm der öfo CUSO4 + 51100 Lösung gebracht und nach 48 Stunden langem Contact abfiltriert. Im Filtrat wurde das Kupfer bestimmt. 200 ccm der ursprünglichen 5^^ Lösung enthalten 3.482g CuO. dieselbe Menge vom Filtrat enthält 0.888g CuO. daher absorbierte die Gartenerde 2.594 9^ CuO. oder 2.068^/. Ca. Meine Versuch ergaben, dass 2 Topfpflanzen von Phius Tliunhergii nach 4 Monaten noch lebendig waren, selbst als 100 ccm der bYo Lösung auf 12 Älal (d. h. eine gesammte Menge ca. 17 gr Cu) gegeben worden waren. Das Trockengewicht der Topferde betrug ca. 700g daher müssen ca. 15 g Cu schon von der Erde absorbiert gewesen sein während die überschüssige Menge des Kupfers theils noch in den Erde Idieb, theils aber durch späteres Nachgiessen entweder vom Wasser ausgewaschen wurde oder theilweise in die Pflanze eindringen konnte. Indessen war noch kein erheblicher nachtheiliger Einfluss auf den oberen Tlieil der Pflanze wahrzunehmen. Bei einer der obenerwähnten Topfpflanzen, die am 2 Juni von der Erde befreit \vinxlen, konstatierte ich, dass die Wurzeln mit Ausnahme von denjenigen welche in der Mitte der Topferde lagen fast vollständig abgestorben waren. Nachher wurde der aufgenommene Kupfergehalt sowohl der Blätter als auch des Stengels quantitativ bestimmt : — 382 H. HATTORI : EINWIRKUNG DES 1. Trockeno'cwicht der Blätter 36.9o;. Cii-Gehalt derselben O.OOOlop;. 2. Trocken erewicht des Sten2;els 3o.8o\ Cu-Gehalt derselben 0.00055g. Vergleicht man diesen Gelialt an Kupfer mit demjeingen welcher in Boden l)leibt, so sieht man, dass der erstere nur ein Bruchstück des letzteren ist. Natürlich wird das Kupfer beim Uebergiessen der Erde nicht gleichmässig im Boden absorbiert, die oberflächlichen Erdsichten empfangen zuerst eine beträchtliche INfenge, während die nächst tiefere Schicht noch von dem Metall frei bleibt. Bei fortdauern- der Berieselung aber wird der Kupfergehalt des Bodens allmählich von oben nach unten fort schreiten. So werden die Wurzeln an der oberen Erdschicht zuerst Schädigung erfohren und abgetötet werden bis schliesslich der nachtheilige Einfluss auf das ganze Wurzelsystem verbreitet wird. So lange der oberirdische Theil der Pflanzen noch lebendig bleibt, wenn auch ihre Wurzeln schon abgestorben sind, muss der Transspirationswasserstrom durch solche leblose Wur- zeln stattfinden. Auf diesem Gebiete, hatten Ha n s e n^) und Jans e") bereits festgestellt, dass die PHnnzen nach Tödtung der Wurzeln längere Zeit vollkommen frisch l)leibcn und eine beträchtliche Wassermenge mit Hülfe der abgestoi-benen Wurzeln aufnehmen können. Die Zellen der lel)enden Wurzeln besitzen einen ei'heblichen Widerstand für Permeabilität der gelösten Stoffe. Tötet man aber die Zellen durch giftige Stoffe, so können nicht nur diese Stoffe sondern auch alle gelösten Substanzen leicht durch die 1) Hansen, Ein Iioitrag znr Kenntniss des Transpiriitionswasseistrnnies. Arb. d. bot. Inst, in Würzburg. lîd. Ill, p. 808 n. IÎI3. 2) J anse. Die Mitwirkung der Markstralden bei der Wasserboweguug im Holze. Pringsh. Jalirb. f. wiss. Hot., Bd. XVIII, p. 17. KUPFERSULFATS AUF EINIGE PFLANZEN. Oöo Wurzelzellen liinclurcli dringen') nnd werden mit dem Trans- spirationsstrom anfwärts steigen. Auf diese Weise nehmen die obengenannten Topfpflanzen, die mit stärkeren Kupfer sul fatlösung begossen sind, doch noch eine geringe Menge des Kupfers durch die abgestorbene Wurzel ins Körperinnere auf und kann sogar eine nicht unbedeutende Anhäufiino; im demselben stattfinden. V. Die Abhängigkeit der Einwirkung des Kupfersulfats von der Luftfeuchtigkeit. Natürlich ist die Transspiration der Pflanzen von äusseren Bedingungen, besonders von Feuchtigkeit und Temperatur der Luft, abhängig") und erhebliche Verminderung derselben ist unvermeid- lich wenn die umgebende Luft mit Dampf gesättigt ist. Demnach muss die Einwirkung der Giftlösung, welche mit dem Trans- spirationsstrom in den Pflanzenkörper eindringen kann, durch Luftfeuchtigkeit mehr oder minder beeinflusst werden. Nobbe") bemerkte, dass die vergifteten Pflanzen längere Zeit turgescent bleiben können und die Giftwirkung nicht auf- gehoben wurde, wenn die Pflanzen im feuchten Räume oder im Dunkeln gehalten worden sind, ferner zeigte sich bei Versuchen Gaunersdorfers'') mit Lithiumsalzen eine ähnliche Thatsache. Um die Einwirkung des Kupfervitriols unter der erwähnten 1) Str asbu r ger, I.e. p. 852-853 und dort citierte Arbeit von Saussure, Eeclierches chimique sur la végétation 1804. 2)Pfefl'er, 1. c. p. 221 u. 227. 3) Nobbe, Untersuchung über die Giftwirkung der Arsens, Blei und Zink &c. Landwirth. Vers. St. Bd. XXX, 1883. 4) Gauner sdor fer. Die Verhalten der Pflanzen bei Vergiftungen speciell durcli Lithiumsalze. Landw. Vers. St. Bd. XXXIV, 1887, p. 193. 384 H. HATTORI : EINWIRKUNG DES Bedingung festzustellen, wurden die Halme von Gerste und die Triebe von Bohnen in 0.1'^ Kupfervitriollösung eingetauclit und ein Theil der Kultur in einem dampfgesättigten glasbedeckten Kasten der andere offen im Zimmer unter gleicher Temperatur «•ehalten. Eine Anzahl diesbezüglicher Versuche ergab, dass der Trans- spirationsstrom, wie erwartet, auf die Giftwirkung der Kupferlösung einen wesentlichen Einfluss übt und die Beschädigung des oberen Theils der Pflanzenkörper durch die Luftfeuchtigkeit erheblich vermindert wird. In unseren Versuchen blieb z. B. die Bohne 7 TaQ;e und die Gerste 4 Tage lano; im sehr feuchten Bäume noch gesund, während die Kontrollpflanzen in gewöhnlicher Zimmerluft schon lange abgestorben waren. VI. Das Verhalten einiger Kulturpflanzen in Kupfervitriollösungen. Um die minimale Grenzkonzentration für die Giftwirkung der Kupfervitriollösung zu ermitteln, kamen in unseren Versuchen in aus Glas destilliertem Wasser kultivierte Pisum- und 3Iais- Keimlinge in Anwendung^), und zum Vergleich führte ich auch einige Kulturen mit aus einer Kupferretorte destilliertem AVasser aus. Die Zuwachsgrösse der Kulturen wurde v o r und n a c h dem Versuche durch IMessung der Länge der AVurzeln und der Spros- sen von der Tnsertionsstelle der Kotyledonen aus oder der Endo- 1) Der Grund wanim ich bei den Kulturen keine Nälirlc'wuug li inzugefügte, wurde scliou in Kap. TI erwiihnt. IvUPFERSULFATS AUF EINIGE PFLANZEN. 385 sperms mit einander verglichen und ferner wui-de das gesammte Trockengewicht der (Sprosse und Wurzehi bestimmt. Ich beobachtete bei einigen Versuchen mit Pisum sativum, dass die minimale Konzentration der Kulturlosungen, welche auf die Wurzeln nicht mehr tödtlich einwirkt, zwischen O.OOOOo?» und 0.00001'?^ hegt, ferner konnte ich auch eine Schädigung durch aus einer Kupferretorte destilliertes Wasser bemerken, doch zeigte in einem Falle die Pflanze, welche in solchem Wasser gehalten worden war, nicht nur keine Schädigung sondern eine kräftige Entwickelung und somit keinen Unterschied zu den Kontrolpfianzen. Folgende Tabelle zeigt die Resultate : Konzentration der Lösungen. Die Länge der Sprosse in cm. Mittel aus je 5 Plianzen. Die Länge der Hauptwurzehi in cm. Mittel aus je 5 Pflanzen. -èl « . Verhalten •x^og der Wurzeln c ='^^ , nach dem S * à"' 1 gg.s"^ Versuche. H^ 1 i 1 N 1 Vor und nach dem Vei suche. Zu- wachs. Vor und nach deiu Versuche Zu- wachs. Kontroll. Aus Kupferretorte dest. Wasser. 0.000001?^ CuS0, + 51i.,0. 0.000005 ?ö CuS0, + 5H,0. 0.00001?^ CuS0, + 5H,0. 0.00005?^ CuS0, + 5H,0. 0.0001 f^ CuS0,+5H„_0. 5.5 5.5 5.7 5.4 4.5 5.7 5.3 26.0 14.7 20.4 22.5 17.4 15.0 14.3 20.5 9.2 14.7 17.1 12.9 9.3 9.0 15.9 14.2 13.4 14.5 15.5 14.5 13.6 22.6 15.4 24.0 20.2 18.5 15.1 13.8 6.7 0.220 lebend 1.2 0.101 abgestorben 10,6 0 143 : lebend 5.7 0.132 3.0 1 0.103 i 0.6 ] 0.097 abgestorben 1 0.2 1 0.083 12-23°C. Aus anderen Versuchsreichen mit Mais folgt, dass eine 0.000001 ?ö Lösung einen nachtheiligenden Eintluss auf die Ver- suchspÜanzen nicht mehr ausübt und dass die minimale Konzen- tration, welche auf Wurzelzellen derselben giftig ist, in der That J86 H. HATTOEI : EINWIRKUNG DES zwisclieii 0.000001^ und 0.000005?^ liegt; selbst 0.0000019^ Lösung wirkt auf die Liingenzuwachs der Seiten- sowie Haupt- wurzel ziemlich stark hemmend ein, und vermindert das Trocken- gewicht der Pflanzenkörper. Ebenfalls wirkt das aus einer Kup- ferretorte destillierte Wasser auf die Wurzeln meistens vergiftend. Die Einzelresultate sind in folgender Tabelle zusammengestellt : Konzentnition der Lösungen. Kontroll. Aus Kupferretorle ilcstl. Wasser. O.OOOCOl^o CuS0, + 5H„0. ().0(J0U05?o CuS0,+5H,0. 0.00001?^ CuS0,+5H,0. 0.00005% aiS0,-i-5H„0. 0.0001% CuSO, + ILO. Die Länge der Sprosse in cm. Mittel aus je 5 Ptlanzen. Die Länge der Hauptwurzeln in cm. Mittel aus je Pflanzen. Vor und nach dem Versuch. 7.1 19.6 7.2 16.r, ().5 17.5 7.0 16.7 6.(3 18.0 5.9 12.7 7.4 11.0 Zu- wachs. 12.5 9.1 11.0 9.7 11.4 6.8 o.ü Vor und nach dem Versuch. 13.2 25.1 14.4 14.5 11.9 24.8 14.0 18.8 113 12.4 12.4 13.0 14.6 14.9 Zu- wachs- 11.8 0.1 12.9 4.8 1.1 0.6 0.3 rgg- I Verhalten !5i=^_I « , der Wurzeln | = S£ nach dem N S Versuche. 0.130 ' lebend C.106 I abgestorben 0.115 I lebend 0.109 labgestorben 0.110 0.069 0.062 15-24°C. Dass solche grosse Verdünnung auf Mais und Er1)sc noch schädlich einwirkt, ist meines Wissens noch nicht bekannt. Die Arbeiten von Hasel lioff^) und Otto^) geben keine Auskunft darüber ; die genannten Autoren experimentierten mit viel stärkeren Lösungen. Nach den Versuchen von Heald"') konnte 1/51200 Gr. l)IIaselliofl, I.e. p. 261 und iul' 2) Otto, 1. c. p. 327-334. 3) Ile aid, 1. c. p. 139. KUPFEESULFAÏS AUF EINIGE PFLANZEN. 387 Mol. Kupfersulfatlösung den AVurzelzuwaclis von Plsmri hemmen und eine halb so konzentrierte Lösung übte auf Erbse- und ]\Iais- Wurzeln nicht mehr eine nachtheilige Einwirkung aus. Diese Versuche He aid 's sind a])er auf eine kurze Zeitdauer beschränkt und so konnte natürlich irgend ein Nachtheil, der bei längerer Einwirkung hervortreten könnte, nicht wahrgenommen werden. In unseren Versuchen wirkte eine weitaus verdüntere Lösung als 1/102400 Gr. I\Iol. nach 2 Tagen bei Mais und in 3-10 Tagen bei Erbse sicher schädigend ein. Auch das aus Kupfergefössen destillierte Wasser übte bei meinen Versuchen auf beide Pflanzen mit Ausnahme von einem Falle sicher eine beeinträchtigende Einwirkung aus. Bekanntlich erklärte Naegeli^) das Vergiften der Spivogyra Fäden durch destilliertes Wasser durch die Annahme einer oligodynamischen Wirkung des Kupfers. Loew") beobachtete in der That einen geringen Kup- fergehalt in aus Kupfergefässen destilliertem Wasser. Auch die Versuche Otto's') ergaben eine schädliche Wirkung solchen Wassers beim Weizen. Was die Symptome an den erkrankten Wurzeln anbelangt, so ist zu bemerken, dass ihr Aussehen zuerst milchweiss wird, dann von Scheiteltheil beginnend sie ihrer ganzen Länge nach ge- bräunt werden, dann der Turgor verschwindet und die Wurzel abstirbt. In den selbst in sehr verdünnte Kupferlösungen ein- gestellten Wurzeln trat gewöhnlich eine bläuliche Färbung an 1) V. Naegeli, I.e. 2)Loe\v, Bemerkung über die Giftwirkung des destillierlen Wassers. Landw. Jahrb. Bd. XX, 1891. o)Otto, I.e. p. 326. Auch die Ionen- Theorie wurde verwendet um die Giftwirkung der Ku])ferbalze zu erklären (Copehuul und Kahlenberg, The Influence of the Presence of l>ure Metalls uiwu Plants. Trans, of Wise. Acad. Vol. XII, 1899.) 388 H. HATTORI : EINWIRKUNG DES der Waclisthumszoïie oder den beiiachbcirten Gevvebetlieilc ein. Diese Färbung ist durcb das aufgespeicherte Kupfer verursacht, was mit FerrocyankaUum leicht demonstriert werden kann. Können die lebenden Wurzeln, ohne Schaden verdünnte Kupferlösung aufnehmen ? Darüber sind die Beobachtungen nicht übereinstimmend; so war De Can do lie ^) der Ansicht, dass Kupfer von den Pflanzen aufgenommen werden kann, ferner ge- langten Phillips'), Freytag'^), Tschirch') und andere zu glei- chem Ergebnisse. Anderseits scheint es nach Otto'^) nicht der Fall zu sein. In neuerer Zeit, äusserte Overton'"') die Ansicht, dass ,,alle Verbindungen, welche schon in massig verdünnter Lösung zum grössten Theil in die Ionen zerfallen sind, nicht merklich in das Protoplasma eindringen, so lange die Grenzschichten des Protoplasts unbeschädigt sind," und ferner ,, durch eine aktive Pesorption kön- nen noch diese Substanzen unter gewissen von der Lebensthätigkeit der Zelle abhängigem Umständen, von den Zellen aufgenommen werden." Unzweifelhaft ist aber wie Pfeffer') betonte, dass „die Pflanze sehr erhebliche Mengen von sehr giftigen Körpern sogar in gelöster Form speichern kann, wenn nur durch Darbietung einer genügend verdünnten Lösung dafür gesorgt ist, dass in dem lebendigen Protoplasmaleib nie eine schädigende Konzentration erreicht wird." So ist es höchst wahrscheinlich dass in unseren Versuchen, eine so verdünnte Kupfervitriollösung wie von 1) De Candollc, Pliysiologic vc'gélale. Ed. I, 183"2, ]>. 'JS9. 2) Phillips, I.e. :;) Frey tag, I.e. 'i) Tscliircli, le. \>. 17, 5) Otto, I.e. p. ;;;;4. 6) Overton, Tehor die osinotiselie Ei^enseliaft der Zelle in ihrer Ijedeuluni; für die Toxieologie luid l'h:iraial\t)logie. Sind. AI id., IS'.Hi, p. 10. 7) Pfeffer, I.e. ii. 429. KUPFERSULFATS AUF EINIGE PFLANZEN. 389 0.000001 '^ö in die lebenden AVurzcln von Mais und Erbse ohne Schaden eindringen kann, ^^'äh^end eine konzentriertere Lösung ihren Weg ins Zellinnere erst finden kann wenn die Lebensthätig- keit des Zellleibes beeinträchtigt worden ist. Um eio'cne Erfahruno; über die relative Schnellia'keit des Ein- dringen« der Cu-Lösung in lebende resp. abgestorbene Wurzeln zu gewinnen, hatte ich einerseits die mit warmem Wasser getödteten Wurzeln von Fisuni und anderseits lebende Wurzeln in 1 Yo Kupfer- vitriol-Lösung gestellt. Schon nach einer Minute, zeigten die todten Wurzeln der ganzen Länge nach, besonders an der AVachs- thumszone, starke Kupferreaktion, während die lebenden Wurzeln noch ungefärbt blieben. Erst nach drei Stunden, trat bei den gesunden Wurzeln in den um den Scheitel liegenden Theilen und in den Seitenwurzeln eine ebenso intensive Kupferreaktion ein, wie bei deu getödteten Wurzeln schon nach einigen Minuten. Die Versuche zeisren «"cnücrend wie schwer eine mässis; konzentrierte Giftlösung in lebende Zellen eindringt, da sie einen gewissen Wider- stand seitens des Zellleibes zu überwinden hat ; ganz anders ist es aber bei todten Zellen wo das Eindringen nur auf mechanische Weise stattfindet. Eine Bemerkung ueber die Desorganisationserscheinungen DER Würzelzelle. Die Desorganisationserscheinungen der Spyrogyra-ZoWew in Kupferlösungen wurden von Naegeli^) eingehend studiert, welcher 1) V. Naegeli, I.e. p. 33. 390 H. HATTORI : EINWIRKUNG DES zeigte, (lass die löslichen Stoffe nach ihrer Konzentration drei Arten tödtlicher Erkrankung hervorbringen ; nämlich in grösster Menge des Kupfers die ^''^'^ysik^üsche, in massiger Menge die chemische, in geringster Menge die oligod3mamisclie Todesart.^) Diesbezügliche Versuche wurden auch in neuerer Zeit durcli Cramer^) und Kumm^) angestellt. Nach den letztgenannten zwei Autoren sollen die oligodynamische Erkrankungen der Spiro- ffyra-ZeWen erst von 0.001% Verdünnung des Kupfersulfets an auf dem Objectträger auftreten. Es wurde aber in unseren Versuchen mit den Wurzelzellen von Mais und Erbse nicht festgestellt, dass in einer gleich starken solchen Lösung eine el^cn solches Todessymptom wie Naegeli angab stets zu Stande kommt. Auch in einer stark konzentrierten Lösung (z. B. 10 fo) findet nach der Untersuchung von Klemm'') das Absterben der Momordica-ZeWe ohne jede Configurationsänderungen statt und die Kontraktion des Pasmas ist nach demselben Autor nicht ein specifisches Todessymptom, sondern nur die Folge der Einwirkung des schädigenden IMittels in geringeren Grade. Obgleich die in unseren Versuchen verwendeten Kupferlösung- en (0.000019^, 0.00019^, 0.0019^, O.Ol 9^, OÄ^/o und l^/o) sowie das aus einer Kupferretorte destillierte Wasser besonderes charakteris- tische Desorganisationserscheinungen an den Wurzelzellen nicht zeigten, so sind doch folgende Aenderungen wahrzunehmen: das 1) Nach Loew's Ansicht heniht die oharaktcristiwclie Wirkmia; lioch verdünnter Knpfcr- l()snngen auf Splrogyra darauf, dass die Ciiloropliyllspirale Kuiifcr s|)eichcrt dalier lantije vor dem Nucleus und Cytoplasnia alistirbt und sich allein contrahiert. (Die chemi^:clle Energie der lebenden Zellen, 1899.) 2) Cramer, N.achtrag zu Naegeli's Untersnchnngon iil)er oligodynaniisclie Erscliein- nngen &c., 1893, p. 43. 3) Rnmm, Zur Kenntnifsdor "Wirkimg der T5 »rdeanx-Biiilie und ilirer IVstandtheile auf Spirnyyra Jotxjnla und di'r Ureddsporen von Perononpora cornncda. Fünfslück, Zeit. f. wiss. Bot. P.d. I, 1895, p. 97. 4) Klemm, Desnrganisationsensohcinnng der Zelle. Pringsh. Jahrb. f. wiss. Pot., Pd. XXVIII, 1895, p. 670. KUPFERSULFATS AUF EINIGE PFLANZEN. 391 Plasma zieht sich mehr oder minder von der Membran in nnregel- mässigen Umrissen zurück und wird sehwaeli dunkel geßirbt, der Zellsaft wird trül), und der Kern schrumpft einseitig zusammen. VII. Kupfervitriol als Wachsthum beschleunigendes Reizmittel auf Pilze. Es erschien mir interessant die Keizwirkung des Kupfers bei 2 Pilzarten {Aspergillus 7iiger und Pénicillium glaucum) zu unter- suchen und obgleich meine diesbezüglichen Versuche nicht zahl- reich sind beweisen sie doch, dass das Kupfer, wie viele andere Metalle, auch eine beschleunigende Einwirkung auf das Wachsthum genannter Pilze ausüben, und die Ernte der Pflanze bedeutend vergrössern kann ; so betrug z. B. bei Aspergillus welcher in einer 0.004% Kupfervitriol haltigen Nährlösung kultiviert wurde, in einem Falle das Erntegewicht 0.983 g. während bei Normal- kultur nur 0.489 g. Bei Pénicillium Avurde in 0.008% Kupfer enthaltenden Kulturflüssigkeit 0.969 g. Ernte erhalten, während in der nicht kupferhaltigen Lösung nur 0.740 g. In fast allen Kulturen in Pfeffers Lösung erreichte das Mycelium beider Pilzarten eine sehr beträchtliche Entwickelung und war mit einer reichlichen Menge von Calciumoxalat-Krystallen, welche durch die Verbindung der von Pilze ausgeschiedene Oxalsäure mit Ca-Salz der Nährlösung gebildet war, dicht bedeckt. Die Conidienbildung von Aspergillus trat bei allen Kulturen fast gleichzeitig ein, während bei Pénicillium mit steigenden Kon- zentrationen sie immer langsamer stattfand. 392 H. HATTORI : EINWIRKUNG DES Abgesehen von kleinen Schwankungen ist die 0]itimale Kon- zentration bei dem ersteren Pilz ca. 0.0049^ und beim letzteren ea. 0.008«/.^). VIII. Resume der Resultate. 1. Die Erkrankungssymptome eines Nadelholzzweiges, der in einer sehr verdünnten Kupfervitriollösung verweilte, sind folgende : der Siebtheil erhält zuerst eine gelb bräunliche Verfärbung, die Chlorophyllkörper sind misgestaltet und schliesslich tritt Bräu- nung der Nadeln ein. Die Verfärbung schreitet nun von unten nach oben fort und zuletzt verbreitet sie sich auf alle Theile des Zweiges. 2. Die minimale Konzentration des Kupfervitriols, welche auf Zweige von Cryptomeria^ Pinus und Thuja schon schädlich ein- wirken kann, liegt zwischen 0.001-0.0059^. Thuja ist etwas wider- standfähiger als die zwei anderen Arten. 3. Die Gartenerde besitzt eine merkliche Al^sorptionskraft für Kupfersalze und demgemäss dient sie in ihm erwachsenen Pflanzen als ein entgiftendes Mittel, so dass stark gekupferte Topfpflanzen auf längere Zeitdauer ihre Lebensthätigkeit fortsetzen können. 4. Die Giftwirkung des Kupfersalzes ist von der Luftfeuchtig- keit abhängig, insofern diese die Grösse des Transpirationsstromes beeinflusst. 5. Die Wurzeln von Erbe und von Mais sind gegen das Kupfer so empfindlich, dass sie schon in stark verdünnten Kupfer- 1) Vergl. Ono, Ueher die Waclistlmmbcsolilonnigung einiger Algen nnd Pilze (lurch chemische Hei//'. Jonrn. Coli. Science, Ini]i. Univ., Tokyo. Vol. XIIT, \W\ \t. 102, 179 u. 180. KUPFEKSULFATS AUF EINIGE l'FLANZEN. oDo vitriollüsuiigeii absterben. Am empündliclisteii ist gewöhnlich die Waehsthumszone. Die erkrankte Wurzel wird zuerst niilch- weiss dann schwach gelblich braun, und schlieslich dunkelbraun. 6. Die minimale Konzentration der Kupfervitriollösuug, in welcher die Erbsenwurzeln lebendig bleiben können, liegt zwischen 0.00005^^-0.00001^ und bei Maiswurzeln zwischen 0.0000059^,-0.0000019^. Obschon eine 0.000019^ Lösung auf die Wurzeln von Erbse und eine 0.0000019^ auf diejenigen von Mais nicht mehr tödtlich einwirken, führen sie doch noch einen schädlichen Einfluss auf den Zuwachs derselben herbei. 7. In Uebereinstimmung mit früheren Angaben kann das aus Kupfergefässen destillierte Wasser auch eine tödtliche ^Einwirkung auf die Wurzeln hervorrufen. 8. Das Kupfer kann als Keizmittel das Wachsthuni einiger Pilze beschleunigen ; die günstige Konzentration liegt bei Pénicillium bei ca. 0.008?^ und die bei Aspergillus bei ca. 0.0049^. Zum Schluss sei es mir erlaubt, Herrn Prof. Dr. Miyoshi auf dessen Vorschlag ich die vorliegenden Studien unternahm, für seine vielfache Anregung und Unterstützung den verbindlichsten Dank aussprechen und auch Herrn Prof. Dr. Matsumura sage ich an dieser Stelle für das wohlwollende Interesse welches er meiner Arbeit entgegengebracht hat, meinen herzlichen Dank. Juni, 1900. Botanisches Institut, Kaiserl. Universität zu Tokyo. 394 H. HATTOEI. Inhalt. I. Einleitung und Litteratur. II. Methodisches. III. Das Verlialten abgesclinittener Zweige gegen Kupfersulfatlösungen. IV. Das Verhalten von Topfpflanzen gegen Kupfersulfotlösungen. V. Die Abhängigkeit der Einwirkung des Kupfersulfats von der Luftfeuchtigkeit. VI. Das Verhalten einiger Kulturpflanzen in Kupfersulfatlösungen. VII. "Kupfervitriol als Wachsthum beschleunigendes Reizmittel auf Pilze. VIII. Eesumc der Eesultate. TAFEL XIX. Wasserkiilturcn vua Fisum mUvum (phutogruphiert am Ende des Versuclies). 1. Ohne Zusatz von Kupfervitriol. Kultur in aus Glasgefüssen destilliertem Wasser. 2. Ohne Zusatz von Kupfervitriol. Kultur in aus einem Kupfergefiiss destil. Wasser. 0. Mit Zusatz von 0.000001^ CuS0^+5H„0. 4. Mit Zusatz von 0.000005^ „ 5. Mit Zusatz von O.OüOOl^ „ 6. Mit Zusatz von 0.00005^ „ 7. Mit Zusatz von O.OOOI^ „ (vergl. S. 385.) Jour. Sc. Coll. Vol. XV, PI. XIX. Anatomische Studien über wichtige Faserpflanzen Japans mit besonderer Berücksichtigung der Bastzellen. vox K. Saito, RigalîusU. Mit Tafeln XX u. XXI. I. Einleitung. Die vorliegenden üntersucliimgen verdanken ihren Ursprung dem Umstände, dass es trotz zalilreiclien, zur Gewinnung der Gespinnst- und anderen teclmiscli nötliigen Fasern dienenden ein- heimischen und kultivierten Pflanzen Japans eine einschlägige Litteratur bislang nicht existiert und somit eine Bearbeitung der Anatomie des Basttheils unserer Faserpflanzen unter Berücksichti- gung der technischen Anwendung dringend nöthig ist. Ich verzichtete auf eine monographische Aufzählung aller in Frage kommender Pflanzen, aber bestrebte mich, erstens eine allgemeine anatomische Charakteristik, zweitens die Anordnung im Pflanzen- körper und drittens, so weit möglich, die Entwickelungsgeschichte der Bastfosern unserer Objektpflanzen klar zu stellen. 396 K. SAITO : ANATOMISCHE STUDIEN Was die Terminologie nnbetrifft, so scliliesse icli mich über- haupt Haberlandt an, und verstehe unter „Bast" nicht allein die speeifisch-mechanischen Fasergewebe in der Kinde des Dicoty- lenstammes, sondern auch das entsprechende Gewebe in Monocoty- len und diejenigen in den interxylären Phlœm der Dicotylen. Der Ausdrück ,, Faser " ist bekanntlich auf höchst verschied- enen Zellformen^) verwendet ; so verstehen wir darunter Bast- und Holzfasern dicotyler und monocotyler Pflanzen, Gefössbündel der Blätter, Pflanzenhaare u. s. w. In der vorliegenden Arbeit sind hauptsächlich nur Bastfasern in Betracht gekommen, obgleich der Unterschied zwischen Bast- und Holzfasern nicht so sehr auf mor- phologischen Merkmalen, sondern vielmehr in topographischer Lagerung liegt. Um Wiederholung zu vermeiden, wird eine allgemeine Be- sprechung der Litteratur hier nicht unternommen werden : man findet die Litteraturangabe an den passenden Orten der folgenden Abschnitte. Angewandte Reagentien. 1) Jod und Schwefelsäure. Ich stellte Jodlösung nach Höhnel") folgendermassen her: man löst 1 gram Kaliumjodid in 100 gram destilliertem Wasser und setzt einen Ueberschuss von Jod zu, bis die Lösung dadurch gesättigt ist. Die angewandte Schwefelsäure^) besteht aus 2 Volumentheilen 1)C. R. Dotlge, A descriptive catalogue of useful fiber plants of the world. 1897. 2) F. Höhnel, Mikroskopie der technisch verwendeten Faserstoffe. 1887. p. 21. 3) Bert hold, Ueber die mikroskopischen Merkmale der wichtigsten l'llanzen fasern. 1883. (Ref. in Just's Botanische Jahresberichte.) UEBER WICHTIGE FASERPFLANZEN JAPANS ETC. 397 reinsten Glycerin, 1 Volnmentlieil destillierten Wassers und 3 Volnmentheilen concentrierter Schwefelsäure, Behandelt man die Bastzellen mit den beiden Reagentien, so ßirben sich die aus reiner Cellulose bestehenden Zellwände rein blau, während bei den mit anderem Substanz incrustierten eine gelbe oder grüne Färbung eintritt. 2) Chlorzinkjod. Die Lösung^) wird leicht hergestellt durch Auflösen von 20 gram Zinkchlorid, 6.5 gram Kaliumjodid, und 1.3 gram Jod in 10.5 c.c. Wasser ; sie färbt reine Cellulose röthlich bis blauviolett und die verholzten Fasern gelb. 3) Phloroglucin und Salzsäure. Behandelt man eine verholzte Zelle mit concentrierter alkoholischer Lösung von Phloroglucin, und fügt nach Trocknen des Objektes einige Tropfen concentrierter Salzsäure hinzu, so tritt eine intensive Rothflirbung auf. 4) Schwefelsaures Anilin. Eine concentrierte wasserige Lösung färbt die verholzten Zellwände gelb. Die Eeaktions- intensitäten mit Phloroglucin-Salzsänre und schwefelsaurem Anilin stimmten überein. 5) Kupfer ox yd ammo ni a k. Dieses Reagens bereitet man dadurch, dass man kleine Stückchen von Kupferdrehspänen mit Ammoniak übergiesst und dann in offener Flasche stehen lässt. Es löst die, aus reiner Cellulose bestehenden Bastzellen, während die verholzten nur aufquellen. 6) Zinnchlorür. Sättigt man concentrierte Salzsäure mit metallischem Zinn, so entstellt eine Lösung von Zinnchlorür. 1) W. Behrens, Tabellen zum Gebranch bei uiikroskopischen Arbeiten. Dritte Auflage. 1898. p. 147. 398 K. SAITO : ANATOMISCHE STUDIEN Dieses Reagens dient zur Abspaltung des Hadromals^) von verholzten Membranen. 7) M ill on 's eh es Reagens^). Manclie monocotylc Bast- zellen färben sich mit diesem Reagens ziegelroth wie Eiweissstoffe. 8) Alk an na''). Eine concentrierte alkoholische Lösung wurde zum Nachweise des Fettes angewendet. 9) Mazerationsgemisch''). a) verdünnte Chromsäure. b) verdünnte Kalilauge. c)Schulze's Mazerationsgemisch. Manche Bastzellen sind leicht isolierbar, ohne vorausgängige Behandlung mit irgend einem Mazerationsmittel. Nach Isolierung der Bast- zellen wurden die Länge und Breite^) aus einer genügend grossen Reihe von Versuchen genau gemessen. 10) Fixierungsmittel. Ich wandte Flcmming'sche Lö- sung^) mit gutem Effekte an. 11) Farbstoffe"). Fuchsin- Jodgrün ; Boehmer's Haema- toxylin; Methylgrün; Safranin ; Gentianaviolett ; Orange; Fuchsin; Congoroth u. s. w. 1) F. Czapek, Ueber die sogenannten Ligninrcaktion des ILilzcs. Sep.-Alul. ans Hoppe- Seyl er 's Zeitschrift für physiologische Chemie. Bd. XXVII. Heft. 1 nnd 2. 1899. 2) A. Zimmermann, Die botanische Mikroteclmik. 1892. p. 220. 3) Ebenda, p. 69. 4) Ebenda, p. G. f>) Unter Breite ist stets der Dnrchmesser der Faser an ilirer breitesten Stellen zn ver- stehen. 6) A . Zimmer m a n n , I.e. p. 69. 7) Ebenda. UEBER WICHTIGE FASERPFLANZEN JAPANS ETC. 399 UeBERSICHT der UNTERSUCHTEN FASERPFLANZEN MIT IHREN TECHNISCHEN ANWENDUNGEN. Pflanzennamen. Fundorte. (-=M) Gespinnststofl' und seile. Boehmeria niven, Hook, et Ani.^) (Nom. jap. Karamush i). Ganzes Japan. Stengel. Gespinnststofl! B. spicata, Thunb. (Norn. jap. Koakaso). )) )) )» Urtica Thunbergiana, S. et Z. (Nom. jap. Irakusa). )) )> Zwirn, Seile, 9. Leguminocse. Pueraria Thunbergiana, Bentli.'') (Norn. jap. Kudu). Ganzes Japan. Stengel. Gespinnststofl". Wistaria chinejisis, S. et Z.») (Nom. jap. Fuji). !) >) )i 10. Linacese. JÂnum usitatissimum, L.") (Nom. jap. Ama). Ganzes Japan.* Stengel. Gespinnststofl'. 11. Celastracege. Cdantrus articulalus, Thunb. (Norn. jap. Tsuruumemodoki). Ganzes Japan. Stengel. Zwirn, Netze. 12. Vitacese. Viià Coigneliœ, Pull. (Nom. jap. Yamabudo). Ganzes Japan. Stengel. Seile, Sack, u.a. 13. Tiliacese. Corchoms capsular is, L.') (Nom. jap. Tsunaso). Ganzes Japan.* Stengel. Seile, selten Gespinnst. 1) Wiesner, Technische Mikroskopie, 1867. p. 229.; Kohstofte. 187o. p. 458.; Ilöhnel, I.e. p. 46.; Wiesner, Die mikroskopische Untersuchungen des Papiers, etc. 1887. p. ol.; Dodge, /.c. p. 98. u. s. w. 2) Reissek, Die Fasergewebe des Leines, Hanfes, der Nessel, etc. Denkschrift d. Wien. Akad. 1852.; Schacht, Die Prüfung der im Handel vorkommenden Gewebe, etc. 1853. p. 25.; Wiesner, I.e. 1867. p. 110.; I.e. 1873. p. 372.;HöhneI, I.e. p. 36.; Schlesinger, I.e. p. 19.; Pocke, Mikroskopische Untersuchungen etc. Arcliiv der Pharmacie. 1886. p. 613.; W^iesncr, I.e. 1887. p. 26.; Dodge. I.e. p. 106. u. s. w. 3) Schacht, I.e. p. 26.; Wies ne r, /-f. 1873. p. 3S6.; Sch Icsinger, ;.c. p. 20.; Ilolinel, I.e. p. 42.; Focke, I.e. p. 612.; Dodge, I.e. p. 85. u. s. w. 4) Dodge, I.e. p. 275. 5) Ebenda, p. 328. 6) Reissek, I.e.; Schacht, I.e. p. 21.; Wiesner, I.e. 1867. p. 108.; I.e. 1873. p. .359.; Schlesinger, le. y. 26.; Focke, Le. p. 610.; Ilöhnel, I.e. p. 34.; W^iesner, I.e. 1887. p. 26.; Dodge, I.e. p. 219. u. s. w. 7) Wiesner, Indisciie Faserpflanzen. 1870.; /.c. 1873. p. 393.; Sch losi nger , /.c. p. 25.; Focke, ^c. p. 615.; H(»hnel, I.e. p. 43.; Dodge, I.e. p. 125. u. .s. w. UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 401 Pflaiizennamen. Fundorte. (* kultiviert!) Zur Faser- gewinnung verwendbare Tiieile. Technische Anwendung. Tilia cordata, Mill. var. japonica, Miq. (Nom. jap. Sliiiianuki). 14. Malvaceae. Abutilon Avicennce, Gan-tn'.) (Nom. jap. Ichibi). Urena lobata, L. (Nom. jap. 0-bondenkwa). Hibiscus syi-iaciis, L. (Nom. jap. Mukuge). 15. Sterculiacege. Firminia platanifolia, L. (Nom. jap. Aogiri). 16. Thymeleaceœ. Daphne pseudomezereum, A. Gr. (Nom. jap. Onishibari). Edgewoiihia papyrifera, S. et Z.^) (Nom. jap. Mitsumata). Wickstrœmia sikokianum, Fr. et Sav.^) (Nom. jap. Gampi). Ganzes Japan. Ganzes Japan.* Liukiu und Formosa. Ganzes Japan. Ganzes Japan. Ganzes Japan. Südliches Japan.* Südliches Japan. Stengel. Stengel. )j )) Stengel. Stengel. )) Seile. Seile. >j )) Seile. Papier. » II. Die Anordnung der Bastzellen mit ihren histologischen Merkmalen. Seitdem Schwenden er in seinem classisclien Werke über „ das mechanische Princip im anatomischen Bau der Monocotylen "^) zum erstenmale nachgewiesen hatte, dass die lang gestreckten und stark verdickten Bastfasern in dem Pflanzenkörper allein mecha- nischen Zwecken dienen und als ein besonderes Gewebesystem den übrigen Gewebesystemen gegenüber gestellt werden muss, sind 1) Dodge, Le. p. 35. 2) Ebenda, p. 154. 3) Ebenda, p. 327- 4) Leipzig. 1874. 402 K. SAITO : ANATOMISCHE STUDIEN unsere diesbezügiiclien Kenntnisse Dank den weiteren Forseliungen mehrerer Autoren insbesondere von Haberlandt^) bedeutend bereichert worden. Während war im vorliegenden Kapitel die Anordnung der Bastzellen von der Schwend euer -Haber lan dt' sehen Grundlage aus zu betrachten versuchten, so wird doch auch das De Bary'- sche topographische System^) nicht ausser Acht gelassen und genügend berücksichtigt. 1. MONOCOTYLEN. a. Cylindkische Okgane. 1) Bambusa stenostachia. (Fig. 2-4.) Das mechanische Gewebe des Halmes von Baînbusa-ni'ten ist nach Seh wendener 's 14. Tyjxis"') angeordnet. Die Länge der Bastzellen beträgt 0.7-2.8 mm, und die Breite 7-25 /^. Die Dicke einer und derselben Faser nimmt von den Enden nach der Mitte allmählich zu. Man kann hier zweierlei Fasern unterscheiden, nämlich dick- und dünnwandige. Bei den ersteren scheint das Lumen in der Längsansicht nur als eine dunkle Linie oder ein etwas breiterer Streif; und die Zellenden sind schmal und stumpf. Bei den letzteren, d. h. dünnwandigen Bastzellen dagegen sind die Enden breiter und weitlumig. Die Wanddicko ist verschieden ; zuweilen tritt eine dünne Querwand auf, welche selten von einem Tüpfelkanal durchzogen ist. Die !)((. l[;ilK'rl aiiilt, Knl\vickeliin>,'s<;eschidite des iiieclumischcn Gcwcbcsysteius, 1879. 2) De Baiy, Ver^leieluiule AnuU)iiiie. 1877. 8) S eil WC ml euer, I.e. p. Cj. UEBER WICHTIGE FASEEPELANZEN JAPANS ETC. 403 kleinen ovalen Porenkanäle clarclisetzen die ganze Länge der Wandung. In den Querschnitten unterscheiden wir ebenfalls zwei Formen, die eine ist dünn-, die andere dickwandig. Die ersteren sind rund und gross im Umriss, dagegen die zweiten polygonal mit abgerun- deten Ecken, oder ganz rund und kleinzellig. Die dünnwandigen Zellen kommen in Gruj^pen mit deutlichen intercellularen Räu- men vor. Jodlösung färbt die Bastzellen gelb, und auf weiterem Zusatz von Schwefelsäure grünlich. Phloroglucin-Salzsäure färbt die Bastzellen kirschroth. Kupferoxydammoniak färbt die Bastzellen grünlich und bringt sie zur Aufquellung. Millon'sches Keagens färbt die Wanduno; zieo-elroth. 2) Oryza sativa. (Fig. 1, 5, und 0.) Die Diagnose^) S c h w e n d e n e r 's über die Anordnung des Bastes von Gramineen trifft nicht bei Oryza sativa zu, weil, wie von ihm") gezeigt wurde, man bei der letzteren einen conti nuierlichen sub- epidermalen Bastring findet, welcher die kleinen Mestomen der äussersten Reihen umschliesst, und selten mit einigen der tiefer liegenden Gefässbündel verwachsen ist. Die meisten der letzteren sind etwas tiefer ins ]\Iark vergeschoben in gleichen Abständen sowohl von der Oberfläche als auch von einander. Obgleich die Internodiumtheile von den Blattscheiden nicht umhüllt sind, zeigen die Querschnitte dersell^en auch den subepidermalen Bastring, mit welchem die tiefer liegenden Mestomen verwachsen sind. Auf Grund dieser Merkmale sollte Oryza sativa als eine besondere 1) Seil wendener, /.r. p. GO. 2) Ebenda, p. lOG. 404 K. SAITO : ANATOMISCHE STUDIEN Klasse von den übrigen Gramineen unterschieden werden, und zwar mit folgender Diagnose : — Ein subepidermaler continuierlicher Bastring ist vor- handen; subepidermale Bastrippen fehlen gänzlich. Innere Geßissbündel sind meist in grösserer Anzahl in der peri- pherischen Zone des Markes reihenförmig angeordnet, und einige oder alle der Bündel sind mit dem Bastring verwachsen. An den Knoten des Halmes sieht man den subepidermalen Bastring nicht mehr deutlich ; er zeigt einen abweichenden Bau. Unter der Epidermis liegt eine dünne Parenchymschicht, welche unmittelbar an den Bastring sich anschliesst. Auf der Innenseite des Bastrings kommen deutliche Mestomanastomosen vor. Die Länge der Bastzelle beträgt 0.55-1.9 mm, die Breite 4—15 fi. Der ümriss der Bastzellen ist geradlinig, von beiden Enden nach der Mitte in der Breite zunehmend. Die Enden sind schmal oder breit, das Lumen ist thcils linienförmig, theils breit und enthält häufig Plasmareste, selten auch Stärkekörnchen. Die Wand ist glatt, meist 1-2 // dick, und von linksschiefen Poren- spalten durchzogen ; nicht selten kommen dünne Querwände vor. Die Querschnitte sind polygonal, manchmal mit abgerundeten Ecken. Das Lumen ist rund oder oval, und die Wand verschieden dick. Jodlösung gie})t eine gelbe Färbung ; auf weiterem Zusatz von Schwefelsäure wird sie grünlich. Mit Phloroglucin-Salzsäure werden die Bastzellen in den Knoten intensiv, aber in Internodien schwach oder nie roth gefärbt. Von der Wachsthumszone der Internodien aus nach vorliegenden Knoten aufwärts, nimmt die Holzreaktion der Bastzellen allmählich zu; auch auf den Quer- schnitten des Halmes sind die Bastzellen um die Gefassbündel in der Verholzung mehr fortgeschritten. Kupferoxydammoniak ruft UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 405 geringe Aufquellung hervor und färbt grünlich ; die Bastzellen in den Knoten sind schwächer quellbar als in den Internodien. Die Mil Ion 'sehe Reaktion wurde auch bei der Wandung constatiert. 3) Pandaniis odoi'aiissimiis. (Fig. 7 und 8.) Was die Anordnuns; der mechanischen Elemente in der kurzen Stutzwurzel von Pandanus odoratissimus betrifft, so ist sie bereits von Seh wen dene r^) genau beschrieben. Die Lauge der Bastzellen beträgt 0.75-2.15 mm, und die Breite 15-25 //.. Die Breite an einer und derselben Zelle von den Enden nach der IMitte nimmt allmählich zu, aber zuweilen ist der Umriss wellig. Die Wand ist 3-5 /i dick, von linksschiefen Poren- spalten durchzogen. Das Lumen scheint ziemlich breit, selten findet sich eine dünne Querwand. Die Enden sind breit, etwas verdickt, wo das Lumen noch deutlich sichtbar ist. Die Querschnitte sind polygonal, geradlinig begrenzt, mit scharfen Ecken. Die Wand ist gleichmässig verdickt, und das Lumen rund oder oval gestaltet. Jodlösung färbt die Bastzellen gelb, und auf weiterem Zusatz von Schwefelsäure grünlich. Phloroglucin-Salzsäure giebt bei einigen Bastzellen intensive Rothfärbung, bei anderen ist aber die Reaktion eine nur eben sichtbare. Kupferoxydammoniak förbt die Bastzellen bläulich, ohne sie jedoch aufzuquellen. b. Bilaterale Organe. 1) 3Iusa sapientum, var. liukiuensis. (Fig. 14 und 15.) Der Bast in dem Blattstiel von Musa sapientum, var. liukiu- ensis ist nach Schwenden er 's II System^) angeordnet. Die Länge der Bastzellen beträgt 2.65-6.4 mm, meist 3-5.5 1) Scliwendener, Z.c. p. 81. 406 K. SAITO : ANATOMISCHE STUDIEN mm, und die Breite 18-31 ii. Die Bastzellen sind gleiclimässig dick, glatt, mit geringer Wanddicke und haben ein grosses und deutliches Zelllumen, welches nur Spuren von Plasmaresten ent- hält und selten mit dünnen Querwänden versehen ist. Die Enden sind meist gespitzt. Die Querschnitte sind polygonal mit stark abgerundeten Ecken, und schliessen meist dicht aneinander. Das Lumen ist gross, rund ; die Wandung beträgt nur 1-4 ji. Mit Jodlösung werden die Bastzellen gelbbräunlich gefärbt, und auf folgendem Zusatz von Schwefelsäure gelb. Dnrch Phlo- roglucin-Salzsäure werden sie roth gefärbt. Kupferoxydammoniak färbt die Zellen bläulich und lässt unbedeutend aufquellen. Die Mil Ion 'sehe Reaktion wird an der Wandung deutlich erhalten. 2) Agave a7ner'icana. (Fig. 9 und 10.) Der Bast ist an den Blättern von Agave americana wie der vorigen Art nach Seh wenden er 's II System angeordnet. Die Länge der Bastzellen beträgt 0.7-1.9 mm, und die Breite 20-40 fji. Sie sind dünnwandig, glatt, von den Enden nach der Mitte merklich breiter ; stellenweise erscheinen ihre Grenzen durch Anlagerung von umgebenden Parenchymzellen wellenförmig gestaltet. Die Enden sind ausgezogen, und das Lumen ist mehr- mals breiter als die Wandung, mit alhnählich schmaler wer- denden Extremitäten ; die Wand ist sehr dünn und von einem System der linksschiefen Porenspalten durchzogen. Die Querschnitte sind polygonal, geradlinig begrenzt, mit scharfen Ecken dicht aneinander schliessend. Das Lumen scheint rund oder oval, selten etwas eckig, und die Wand ist dicker als bei Musa-fnsern. Jodlösung fiirbt die Bastzellen gelb, und auf weiterem Zusatz UEBEU WICHTIGE FASERPELANZEN JAPANS ETC. 407 von Scliwefelsäure grünlich. Alle Holzreagentien geben die charakteristisclien Färbungen. Knpferoxydammoniak bringt die Bastzellen zu geringer Aufquellung und färbt sie bläulich. 3) Alpinia nutans. (Fig. 11-13.) Die Blattscheide von Alpinia nutans ist eben&lls nach S c h w e n d e n e r 's II System gebaut. F u 1 1 e r e r^) stellte bereits die Anordnung des Bastes in der Blattscheide fest ; ich hatte es deshalb für unnöthig, hier auf weitere Beschreibung einzugehen. Die Länge der Bastzellen beträgt 0.6-2.7 mm, die Breite 10-25 /i. Die Bastzellen sind an beiden Enden schmal aus- gezogen, aber die Enden selbst sind nicht scharf spitzig, sondern etwas abgerundet und verdickt. Die Wand ist 3-ö /^ dick und führt eine Reihe von Längsspalten. Stellenweise erscheint die Zellcontour wellenförmig gestaltet. Das Lumen ist meist breit und enthält oft kernhaltige Plasmamasse. Die Querschnitte sind polygonal mit etwas abgerundeten Ecken. Das Lumen ist breit, rundlich oder oval, und ist die Wand sehr dünn, oder massig dick. Jodlösung färbt die Bastzellen gelblich, und auf weiterem Zusatz von Schwefelsäure grünlich. Phloroglucin-Salzsäure färbt sie rotli und Kupferoxydammoniak grünlich, ohne sie jedoch im mindesten aufzuquellen. Die Millon'sche Reaktion wurde bei der Wand deutlich constatiert. 2. DICOTYLEN. a. Bastbildung in einfacher Ringlage. Diese Gruj)pe bildet einen förmlichen Bastring, der höchstens 1) W. Futterer, Beiträge zur Eut\vicl;elungsgebcliiclite der Zingiberacece. Bot. Central- blatt. B'l. LXVIir. 189G. p. 429. 408 K. SAITO : ANATOMISCHE STUDIEN an einzelnen Stellen kleine Unterbrecliungen zeigt, so class die Bastzellen vereinzelt oder in kleinen Gruppen auftreten. Durcli ausserordentliche Querschnittgrösse zeichnen die Bastzellen sich aus. Hieher gehören : — Boehmeria spicata, Urtica Thunhevgiana^ Linum usitatis- mtium, Celaslrus arliculatus. 1) Boehmeria spicata. (Fig. 42-44.) Die Länge der Bastzelle beträgt 7-26 mm, und die Breite 11-72 A«. Man kann zweierlei Bastzellen unterscheiden, eine mit localen Erweiterungen^), andere ohne solche. Die Wand ist glatt oder deutlich gestreift, und manchmal zeigen die Bastzellen un- gleichmässige Verdickung, so dass das Lumen also verschieden breit ist, und zuweilen ganz verschwindet. Sie enthalten Plasma mit vielen Stärkekörnchen, und treten auch Querwände auf, welche von seitlichen Zellwänden nicht unterbrochen sind. Die Enden sind schmal ausgezogen und etwas abgerundet. Das Zelllumen der Bastzellen, welche die oben erwähnten localen Erweiterungen besitzen, ist anfangs selbstverständlich noch ununterbrochen. Früher oder später kommen aber bei den erweiterten Partien manchmal Einkapselungen des Plasmas durch die Wandlamellen zu Stande, welche also das vorhandene Plasma von den alten Zelllumen völlig abschliessen. Die Dicke der alten Wandlamelle ist bei den erweiterten Stellen dünner als bei den nicht erweiterten. Folgende Messungen zeigen diese Verhältnisse. 1)G. Krabbe, Ein r>citrag zur Kenntniss der Struktur und des Wachsthums der vegetabilischen Zellhilute. Jahrb. f. wiss. l'.ot. Bd. XVI II. 1S87. p. .'346 Kr nennt eine Anzahl local erweiterter ßastzellen, aUr er erwülinte />'. spiiata nicht. UEBER WICHTIGE FASEIIPELANZEN JAPANS ETC. 409 Erweiterte Zellregionen. Nicht erweiterte Zellregionen. Durchmesser. Dicke (1er alten Wandlamelle. Durchmesser. Dicke der alten Wandlamelle. 1 50 /i. 5/.. 18 /i. 8/.. 2 72 „ 4 „ 18 „ 8„ 3 55 „ 5 „ 15 „ 5„ 4 48 „ 7 „ 25 „ 10 „ 5 48 „ 5 „ 24 „ 8„ 6 32 „ 4 „ 17 „ 7„ 7 38 „ 4 „ 19 „ 5„ 8 40 „ 3 „ 11,, 4„ 9 60 „ 6 „ 20 „ 7„ 10 62 „ 4 „ 14 „ 6„ 50.5/z. 4.7/i. 18.1 n. 6.8 IX. Die Querschnitte sind polygonal mit abgerundeten Ecken, oder rundlich. Die Wand ist deutlich geschichtet, und in der Dicke wechselnd. Das Lumen ist meist breit, und mit Inhalt- masse gefüllt. Jodlösung färbt die Bastzellen braunroth, und auf weiterem Zusatz von Scliwefelsäure wird die Färbung himmelblau. Kup- feroxydammoniak förbt die Bastzellen blau und bringt sie schliess- lich zur Auflösung. Holzreagentien geben gar keine Reaktion. 2) Urtica Thunbergiana. (Fig. 57 und 58.) Die Länge der Bastzellen beträgt 5-60 mm, die Breite 20-63 //. Die Breite an einer und derselben Zelle ist ungleichmässig, an den einen Partien schmal, an den anderen bandförmig. Die Bastzellen sind deutlich gestreift und Verschiebungen kommen häufig vor. Das Lumen ist breit und enthält körnige Plasmamassen ; die Enden sind schmal, abgerundet, etwas dickwandig und häufig verzweigt. 410 K. SAITO : ANATOMISCHE STUDIEN Die Quersclniitte sind polygonal mit abgerundeten Ecken, oder oval abgeplattet. Die Wand ist deutlich geschichtet und die Schichten sind manchmal radial gestreift. Das Lumen ist oval oder abgeplattet. Jodlösung fiirbt die Bastzellen braunroth, und auf weiterem Zusatz von Schwefelsäure himmelblau. Durch Kupferoxydam- moniak werden sie blau gefiirbt, und nach starker Aufquellung schliesslich gelöst, mit Ausnahme des inneren Häutchens, welches spiralig gestreift wird. 3) Linuvi usitatissimum. (Fig. 19 und 20.) Die Länge der Bastzellen beträgt 14-85 mm, die Breite 18-25 /^, meist 20 n. Die Bastzellen sind sehr regelmässig ge- staltet, in Breite von den Enden nach der Mitte allmählich zu- nehmend. Die Enden sind konisch zugespitzt, selten stumpf. Das Lumen ist meist zu einer dunklen Linie reduciert, jedoch es kommen Erweiterungen des Lumens, in welchem der Plasmakörper häufig eingekapselt liegt, nicJit selten vor. Der Plasmakörper enthält viele Zellkerne und Stärkekörnchen. Die Wand ist längsstreifig und mit Verschiebungslinien versehen. Die Querschnitte sind rundlich, selten etwas eckig. Das Lumen ist klein, oft punktförmig und mit Plasmamassen gefüllt. Die Wand ist deutlich geschichtet. Jodlösung färbt die Bastzcllen bräunlich, und auf weiterem Zu- satz von Schwefelsäure wird die Färbung himmelblau. Durch Kup- feroxydammoniak wird die Zellwand zuerst stark aufgequollen, und dann gerade oder schief parallel gestreift. Während die Zellwand nach kurzer Zeit allmählich gelöst wird, bleibt die Innenhaut als diuiner, geradgestreckter oder wellig gebogener, selten spiraliger UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 411 Sclilaueli zurück.^) Holzreageiitien geben keine Reak- tion. 4) Celaxtrus arliculahis. (Fig. 37-39.) Die Länge der Bastzellen beträgt 20-70 mm, und die Breite 80-135 /i. Sie sind tangential abgeplattet nnd weitlumig. Auf den taugentialen Schnitten des Stammes scheinen die Bastzellen weitlumig zu sein, wogegen sie auf den radialen Schnitten mit schmalen dunkellinieförmigen Lumen versehen sind. Auf den Längswänden der Bastzellen kommen zwei Streifungssysteme vor, welche zur Längsachse in einem scharfen Winkel geneigt sind. Die Bastzellen besitzten deutliche Verschiebungen und Poren- spalten. Die Enden sind schmal ausgezogen, und an der tangentialen Ansicht erscheinen sie stumpf, wo die Zellwand etwas verdickt ist. Das Lumen ist breit und abgeplattet, es enthält noch eine reich- liche Menge von desorganisierten Plasmamassen. Die Querschnitte sind meist unregelmässig, breit oder abge- plattet. Das Lumen ist auch breit und enthält manchmal gelbe Plasmareste. Die Zellwand ist dick, geschichtet, und von vielen Porenkanälen durchzogen. Jodlösung färbt die Bastzellen braunroth, und auf wT^iterem Zusatz von Schwefelsäure wird diese Färbung himmelblau. Be- handelt man die Bastzellen mit Kupferoxydammoniak, so tritt sich Lösung der Zellmembran sofort ein und am Ende trennt die Innenhaut als spiralig gestreifter SchLuich ab. Die letztere wird mit den in die Porenkanäle der Wand eingelagerten Partien freigelegt. l)Vergl. Wiesner, Technische Mikroskopie 18f)7. p. 108.; Rohstoffe, 1873. p. .371 und Die mikroskopische Untersuchung des Papiers. 1887. p. 28. u. s. w. 412 K. SAITO : ANATOMISCHE STUDIEN Eine ähnliche Thatsache hatte Wiesner^) bei den Bastzellen der Maispflanze einmal constatiert. h) Bastbildung sowohl an der Aussengrenze der PRIM AREN Stränge als auch im Inneren des Sekundarbastes. Der relativen Menge und Vertheilung der späteren Fasern zufolge können dieselbe wie folgt eingetheilt werden. «)^) Einfache llinglage der Bastbündel im ersten Jahre und später bloss einzelne zerstreute oder in kleiner Gruppe auftretende Bastzellen. Der erste Bast besteht aus einem in zahlreiche Bündel aufgelösten Ring ; zwischen je zwei Gruppen verläuft alsdann meist ein Rinden- strahl radial nach aussen oder er stellt mit den benachbarten Sclerei- den^) oder ihren Gruppen eine Verbindung her. — i. e. Tschirch's ,, Gemischter Ring''^) (P/«erar^rt Thunbergimia, Wistaria ehinensis). Hieher gehören : — Boehmeria nivea, ülmus montana, var. laciniata, Puerai'ia Thunhergiana, Wistaria ehinensis. 1) Boehmeria nivea. (Fig. 40 und 41.) Die Länge der Bastzellen beträgt 12.3-24.5 mm, eine für Bastzellen beispiellose Grösse'^). Die Breite misst 40-90 />«, meist etwa 50 /^, und wird nach beiden Enden allmählich schmäler, die Enden selbst sind abgerundet. Bei einigen Zellen ist das Zelllu- men stellenweise ausgefüllt, so dass eine scheinbare Querwand ent- 1) Wies ne r, Mikroskopische Untersnchungen der Maislisclie und der Maisfiiserprodukte. Besonderer Abdruck ans Dingler's Polyt. Journal I. Februarheft 1865, Bd. CLXXV. p. 11. 2)Schwendener, I.e. p. 1 45. .S) A. Tschirch, Beiträge zur Kenntniss des mechan- Gewebesystenis der PHanzen. Jahrb. f. wiss. Bot. Bd. XVI. 1885. p. 308. 4) Ebenda, p. 318. 5) Diese Thatsache stimmt mit den bisherigen Angaben ein. UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 413 stellt. Läiigsstreifungen und Verschiebungen kommen deutlich vor. Die Wand ist 15-30 /^ verdickt und das Trumen enthält desorganisierte Inhaltmasse. Die Quersehnittform ist polygonal, mit abgerundeten Ecken. Die Wand ist sehr deutlich geschichtet und das Lumen breit. Jodlösung färbt die Bastzellen bräunlich ; auf weiterem Zusatz von Schwefelsäure w^ird die Färbung blau, grün oder gelblich. Bei einigen Zellen rufen die Ilolzreagentien die charakterischen Färbungen vor^). In Kupferoxydannnoniak quellen die Bastzellen enorm auf, ohne sich jedoch völlig aufzulösen. 2) Ulmus monlana, var. laciiiiata. (Fig. 31 und 32.) Die Länge der Bastzellen beträgt 1.5-7.5 mm, die Breite 10-20 /><. Die Bastzellen sind sehr dickwandig und deutlich aus zwei Schichten construiert. Die äussere Schicht scheint manchmal von der inneren abgetrennt zu sein. Auch bei den künstlich zerquetschten Bastzellen zeigen sich oft die spiraligen Streifangen der äusseren Schicht, die zuweilen ganz abgeworfen ist, und Verschiebungen der inneren Schicht. Die Enden sind stumpf, und hier erscheinen zwei deutliche Schichten, manchmal die äus- sere von der inneren abgetrennt. Die Querschnittform ist polygonal, dicht aneinanderschlies- send. Die Wand ist dick und aus zwei Schichten construiert. Das Lnmen ist schmal, so dass es nur schwer erkennbar ist. Jodlösnng färbt die Bastzellen bräunlich ; auf weiterem Zu- satz von Schwefelsäure wird die äussere Schicht gelb oder braun, die innere dagegen himmelblau. Die Holzreagentien geben nur mit der äusseren Schicht die charakteristische Färbung. Kupfer- 1) Bei den in Handel kommenden Chinagras (Nom. Jap. Karamushi) findet man nicht mehr die Holzreaktion. 414 K. SAITO : ANATOMISCHE STUDIEN oxydammoniak färbt die Bastzellen bläulich, und bringt sie zu geringer Aufquellung. 3) Pueraria Thunhergiana, (Fig. lG-18.) Die Länge der Bastzellen beträgt 0.95-4.20 mm, die Breite 10-22 /A Sie sind regelmässig gebaut, und verschmälern sich gegen die Enden. Die Enden sind kegelförmig, aber etwas ab- gerundet, selten verzweigt. Das Lumen ist oft sehr schmal und erscheint nur als eine dunkle Linie, enthält manchmal reichliche desorganisierte Plasmamassen. Die Querschnittform ist polygonal, manchmal mit etwas ab- gerundeten Ecken. Die Zellen schliessen sich einander fest an. Die Zell wand besteht aus zwei Schichten, und das Lumen ist verschieden breit, bei einigen Bastzellen erscheint es sogar nur punktförmig. Jodlösung färbt die Bastzellen braun, und auf weiterem Zu- satz von Schwefelsäure wird die äussere Schicht gelb, dagegen die innere himmelblau. Phloroglucin-Salzsäure ruft in der äus- seren Schicht die charakteristische Holzreaktion hervor, während sie bei der inneren ganz ausbleibt. Mit Kupferoxydammoniak behandelt quellen die Bastzellen auf, ohne sich jedoch völlig zu lösen. Das innere Haut erscheint als spiralige oder quergestreifte Schläuche. 4) Wistaria chinensis. (Fig. 85 und 36.) Die Länge der Bnstzellen beträgt 1.3-3.7 mm, die Breite 10-20/^. Die Wand ist, wie bei Pueraria Thunbcrgiana, ganz deutlich in zwei Schichten gesondert. Die Enden sind abgerundet, verdickt und selten verzweigt. Man kann zweierlei Bastzellen unterscheiden, d. h. dick- und dünnwandige. Die ersteren sind UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 415 kleinlumig, wogegen die zweiten, nur selten vorkommenden, durch Querwände in eine Anzahl weitlumiger Kammern getheilt sind, in den Kammern findet sieh Protoplasma mit Zellkernen. Die Wand ist von Porenspalten durchzogen. Die Querschnitte sind polygonal, mit abgerundeten Ecken. Die Wand besteht aus zwei Schichten. Das Lumen führt das Plasma und ist meist breiter als bei Puerarla TImmhergiana. Jodlüsung färbt die Bastzellen bräunlich mit gelber Umran- dung, und auf weiterem Zusatz von Schwefelsäure wird die sekundäre Schicht himmelblau, indem die primäre ganz unverändert bleibt. Jedes Holzreagens giebt die charakteristische Färbung nur mit der äusseren Schicht. Kupferoxydammoniak bringt blaue Färbung hervor und bewirkt Aufquellung der Bastzellwand, wobei das innere Haut ganz ungelöst bleibt. ßf) Einfache Einglage von Bastbündeln in ersten Jahr und noch später stark fortsetzende concentrische Bastbildung. Nach der relativen Anordung der benachbarten Stränge auf dem Querschnitte unterscheidet man zwei topographische Haupt- formen'^). I. Concentrische Schichten von Bastzellgruppen wechseln mit eben- solchen von Weichbast regelmässig ab, und die beiderlei Schichten von benachbarten Strängen passen annähernd, wenn auch nicht immer ganz genau aufeinander. {De Bary). Hieher gehört : — Vitis Coignetiœ. 1) Vitis Coignetiœ. (Fig. 47-49.) Die primären und sekundären Bastzellen unterscheiden sich von einander nicht unwesentlich. Die primären Bastzellen sind 1) Scliwendener, I.e. ^i. 145. 2) De Bary, /.c p. 542. 416 K. SAITO : ANATOMISCHE STUDIEN weitlumig, länger und nicht so reichlich getüpfelt wie die se- kundären. Die Länge der primären Bastzellen beträgt 1-3 mm, bei den sekundären 0.4-0.95 mm. Die Breite beträgt bei den ersten 25-30 />«, bei den sekundären 10-25 n, und nimmt von den Enden nach der Mitte allmählich zu. Die Wand ist gleichmässig ver- dickt, und von linksschiefen Porenspalten durchzogen, aber die letzteren kommen nicht auf den Zellenden vor. Das Lumen erscheint sehr breit, darin kommen 1-4 gerade oder schief gerichtete, dünne unverholzte Querwände vor. Der ümriss ist manchmal wellig contouriert. Die Enden sind schmal, oder abge2:>lattet, manchmal verzweigt. Bei den Querschnittformen lassen sich auch zweierlei Arten unterscheiden. Die primären Bastzellen sind gross, polygonal mit abgerundeten Ecken, aber die sekundären oval und tangential abgeplattet. Bei den letzteren treten die Porenspalten auf den tan- gentialen Wänden reichlicher auf als auf den radialen. Die Wand ist gleichmässig dick und das Lumen oval oder rundlich. Jodlösung färbt die Bastzellen gelb, und auf weiterem Zu- satz von Schwefelsäure grünlich. Phloroglucin-Salzsäure giebt die charakteristische Reaktion, aber mit Salzsäure allein werden die Bastzellen auf dem Querschnitte des Stammes von der cambialen Seite aus allmählich roth gefärbt^). Kupferoxyd am- moniak färbt die Bastzellen blau, indem sie nur wenig aufquellen. II. Concentrische, mit Weichbast abwechselnde Faserzoiien sind zwar im allgemeinen zu imtcr.scheiden, streckungsweise regelmässig angeordnet, im ganzen jedoch unrcgelmässig, indem sie sowohl in einzel- nen Stränj^cn durch Weichbastclemente unlerbroclien als auch in 1) Vergl. A. Tscli ireh, I.e. p. 025. und derselbe, Angewandte l'lhuizenunatoniie. lid. I. 1889. p. 176. UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 417 benachbarten Strängen durch Markstrahlen getrennt und ungleich zahlreich und ungleich vertheilt sind. [De Bari/). Diese Anordnung ist bei dem Baste vieler dicotyler Pflanzen die gewöhnlichste, aber mit mannigfachen Modificationen. Hielier gehören : — Corchorus capsularis, Tilia cordata, var. japoiilca, Abidllou Avicewnœ, Ureiia lobaia, Hibiscus syriacus, Firniinia plalcmi folia, Cannabis saliva, Broussonelia kasinoki, Edcjc- luorlhia papyrifera, Wickslrœmia sikolciamim, Dapime pseiido- mezerewn. 1) Corchorus capsularis. (Fig. 45 und 46.) Die Länge der Bastzellen beträgt 0.6-6.35 mm, die Breite Vö-1'2 n. Die Bastzellen sind in ihrem ganzen Verlauf nur wenig unregelmässig. An jeder isolierten Bastzelle findet man die Verengerung des Lumens (,, Der Nichtparallelismus des äusseren und inneren Contours "^)), Stellenweise erscheint das letztere nur als eine dunkle Linie, die aber nie fehlt. Die Enden sind abgerundet und stark verdickt, häufig weitlumig. Die Querschnitte sind polygonal, geradlinig begrenzt. Die Wand ist verschieden breit, so dass das Lumen verschieden weit erscheint. Jodlösung färbt die Bastzellen goldgelb, und auf weiterem Zusatz von Schwefelsäure dunkelgelb bis braun oder grünlich. Die Holzreagentien geben stets charakteristische Färbung. Kup- feroxydammoniak färbt die Bastzellen bläulich, dann quellen sie etwas auf. 1) Wiesner, Indiaclie Faser pßaiizeii, Sitzungsberichte d. Wien. Akad. 1870, und Eoh- stoflë 1873. p. 399. 418 K. SAITO : ANATOMISCHE STUDIEN 2) Tilia cordaia, var. jccponica. (Fig. 23 unci 24.) Die Länge der Bastzellen beträgt 1.48-2.4 min, die Breite 17-23 jx. Die Breite ist an ein und derselben Faser von den Enden nach der Mitte zu allmählich ver2;rössert. Die Enden sind scharf oder unregelmässig, und erscheint das Lumen als eine dunkle Linie. Die Wand ist stark verdickt, und stellenweise erscheint ihre Urariss wellenförmig gestaltet. Sie ist auch von reihenweise angeordneten Porenspalten durchzogen. Die Querschnittform ist polygonal. Die Zellen sind mit einander fest verbunden, und von den breiten Mittellaniellen umhüllt. Das Lumen ist punktförmig oder etwas weiter. Jodlösung färbt die Bastzellen goldgelb. Diese Farbe ändert sich bei weiterem Zusatz von Schwefelsäure in schmutzigbraun und schliesslich in grünlichblau. Jedes Holzreagens giebt die charakteristische Färbung ; auch beim blossen Zusatz von Salz- säure färben sich auf dem Querschnitte des Stammes die Bast- zellen deutlich rotli. In Kupferoxydammoniak quellen sie unter Bläuung etwas auf; das innere Haut tritt dann sehr deutlich hervor, und die sekundäre Verdickungsschicht lässt schöne Lamellenstruktur erkennen. 3) Firminia platanifolia. (Fig. 27 und 28.) Die Länge der Bastzellen beträgt 1.5-3.0 mm, die Breite 15-20/^. In der Längsansicht scheint die Dicke an einer und derselben Bastzelle ziemlich gleichmässig zu sein, aber häufig kommt dieselbe mit vielen Auswüchsen der Wand vor. Die Wand ist sehr dick und von rundlichen Porenkanälen durch- zogen ; das Lumen erscheint als eine dunkle Linie, und fehlt manchmal ganz. Die mittleren Partien der Bastzellen zeigen selten etwas Anschwellung des Lumens, wo die Zellwäiide relativ UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 419 schwächer verdickt als in die übrigen Stellen. Die Enden sind stets dickwandig, scharf oder etwas abgerundet, manchmal mit Abzweigungen. Auf dem Querschnitte kommen die Bastzellen in Gruppen vor. Sie sind polygonal, mit etwas abgerundeten Ecken. Die Mittellamelle ist breit, und das Lumen sehr schmal, punkt- förmig oder etwas weiter. Jodlösung färbt die Bastzellen goldgelb. Auf weiterem Zusatz von Schwefelsäure werden sie gelbgrün oder blaugrün mit gelber Umrandung, oder bräunlich gefärbt. Jedes Holzreagens giebt die charakteristische Färbung. In Kupferoxydammoniak quillt jede Bastzelle nur schwach unter starker Bläuung auf. 4) Abutilon Avîcennœ. (Fig. 29 und 30.) Die Länge der Bastzellen beträgt 1-2.1 mm, die Breite 8-37 /^. Die Breite der Bastzellen nimmt von den Enden nach den Mittel partien ungleichmässig zu. Das Lumen ist sehr breit, aber verengert sich an manchen Stellen, oft bis zum gänzlichen Verschwinden. Die Enden sind meist etwas abgerundet und verdickt, doch ist das Lumen noch deutlich sichtbar. Die Querschnitte sind polygonal und geradlinig begrenzt oder manchmal etwas abgerundet. Das Lumen ist breit und abgerundet polygonal. Im allgemeinen sind die Querschnitte grösser als die von Corcliorus capsularis. Jodlösung färbt die Bastzellen gelblich, und auf weiterem Zusatz von Schwefelsäure dunkelgelb bis braun. Alle Holzrea- gentien liefern die charakteristischen Färbungen. Durch Kup- feroxydammoniak werden die Bastzellen anfangs blaugrünlich gefärbt und später zu enormer Aufquellung gebracht. 420 K. SATTO : ANATOMISCHE STUDIEN 5) Urena lohata. (Fig. 33 und 34.) Die Länge der Bastzellen beträgt 0.75-2.43 mm, die Breite 14-26 /-«. Die Verdickung der Wände einer und derselben Bast- zelle ist ungleich massig, und kommen Verengerung und Versch- winden des Lumens oft vor. Selten findet man Porenkanäle in der Wandung. Die Enden sind stumpf, etwas verdickt, nicht auffallend breit, und selten verzweigt. Das Lumen ist meist schmal, seltener breit, und stellenweise verschwindet es gänzlich^). Die Querschnitte sind polygonal, mit scharfen oder abge- rundeten Ecken und zeigen eine deutliche breite Mittellamelle, Das Lumen ist sehr schmal, oft punktförmig. Jodlösung färbt die Bastzelleu goldgelb ; auf weiterem Zusatz von Schwefelsäure wird diese Farbe kaum verändert. Alle Holz- reagentien liefern mit Bastzellen die charakteristische Färbung. Kupferoxydammoniak färbt die Zellen ohne Aufquellung blau. 6) Hibiscus syriacus. (Fig. 25 und 26.) Die Länge der Bastzellen beträgt 0.6-1.7 mm, die Breite 12-35 /i. Die Bastzellen sind durch die Kerben und Gesch- längeP) an der Wandung unregelmässig gestaltet. Die Wand ist dünn und von Porenspalten durchzogen. Das Lumen ist weit, selten mit Verengerungen. Die Enden sind schmal ausgezogen, nicht besonders weitlumig, aber häufig verzweigt. In den Bast- zellen kommen häufig Fetttröpfchen vor. Die Querschnitte sind polygonal, geradlinig begrenzt, und au den Ecken etwas abgerundet. Die Mittellamelle ist breit, eben- so das Lumen polygonal mit abgerundeten Ecken. 1) A. Kos oll (Jahresbericlit d. niedeiöstcrreichischen Landoberrealsclmle etc. in Wiener Neustadt. 1894.) sagt, dass die Enden der B:istzellcn von LVcna /o6a^a scharf sinl, und dass ein Verseil winden des Lumens nie vorkommt ; doch diese Angahen stinuiien mit meiner Unter- suchung nicht. 2) Vctilldrd, Etüde sur les fibres végétales textiles. 187G. (Citiert in Dodge, le. p- 191.) UEBER WICHTIGE FASERPFLANZEN JAPANS ETC. 421 Jodlösnng färbt die Bastzellen goldgelb, und auf weiterem Znsatz von Schwefelsäure äiulert sich die Farbe in grünlich. Die Holzreagentien liefern mit den Bastzellen die charakteristische Keaktiou. Kupferoxydammoniak färbt die Bastzellen unter geringer Aufquellung blau. 7) Cannabis sativa. (Fig. 53-56.) Die Länae der Bastzellen beträsft 7-50 mm, die Breite 10-35 ij.. Die Bastzelle nimmt von den Enden nach der Mittel- theil allmählich, aber unregelmässig in der Breite zu. Das Lu- men ist breit, plasmahaltig mit vielen Kernen, und wird nach den Enden hin allmählich verschmälert. Längsstreifungen und deut- liche Verschiebungslinien kommen auf der Wand vor, um welche das Stückchen der Umrandungslamelle oft gehängt ist. Die Enden sind meist abgerundet, sehr dickwandig, manchmal mit seitlichen Verzweigungen^). Die Querschnitte sind polygonal, aber die Ecken abgerundet. Sie schliessen sich an einander stets dicht an. Die Mittellara eile ist sehr breit, die Wand deutlich geschichtet. Das Lumen ist schmal, linienförmig, einfach oder unregelmässig verzweigt. Jodlösung färbt die Bastzellen bräunlich ; auf Zusatz von Schwefelsäure wird diese Färbung himmelblau mit einer gelben Umrandung. Die Holzreagentien geben die charakteristische Fär- bung nur auf der Umrandungsschicht. Trotz der grösseren Breite 1) Das Vorkomiuen der gabeligen Enden bei den Bastzellen von Cannabis mtlva wurde schon von vielen Forschern beobachtet: vergl. Schacht, Prüfung der im Handel vorkommenden Ge- webe. 1852. j). 2G.; Schlesin ger, Examen uiicroscn pique et niicrochimiqne des fibres textiles. 1875. p. 19.; Focke, Mikroskopische Untersuchungen der bekannteren Gcspinnstfasern etc. Archiv, d. Pharmacie. 18SG. p. 19.; Höhnel, Mikroskopie der technisch verwendeten Faser- stofle. 1887. p. 30.; und Wiesner, Die Mikroskopische Untersuchung des Papiers etc. 1S87. p. 29. 422 K. SAITO : ANATOMISCHE STUDIEN der primären Bastzellen, ist die Mittellamella derselben dünner als die der sekundären. Durch Kupferoxydammoniak werden die Bastzellmembranen blau oder blaugrün gefärbt und quellen enorm auf, zugleicli erscheint auf der Wand eine deutliche schiefe Parallelstreifung. Das Innenhäutchen tritt hierbei als spiralig oder quer gestreifter Schlauch auf, welcher breiter als bei der Bastzelle von Linum usitatissimum ist. Wenn aber die äussere verholzte Schicht von der inneren abgetrennt wird, so wird die letztere von Kupferoxydaramoniak ganz aufgelöst, während die äussere Schicht zurückbleibt^). 8) Broussonetia kasinoki (Fig. 21 und 22.) und B. papyo'i- fera. Die Bastzellen von Broussonetia kasinohi sind mit einander locker verbunden, und sind ohne Mazerierung leicht zu isolieren. Die Länge beträgt 1.51-10 mm, die Breite 10-34/^. Die Bastzellen sind mit vielen deutlichen Verschiebungen versehen. Man kann zweierlei Zellen^) unterscheiden, dicke und dünne. Einige der Bastzellen sind dickwandig, und die anderen oft band- förmig flach. Das Lumen ist bei den ersteren linienförmig, und bei den letzteren in der Längsansicht schwer zu erkennen. Bei den bandförmigen Zellen scheinen die Enden breit und abgerun- det, bei den dickwandigen schmälern dagegen scharf. Die Breite der Bastzelle nimmt von den Enden nach der Mitte gleichmäs- sig zu. Sie erscheinen im Längsverlaufe häufig von einer lockeren dünnwandigen Scheide umschlossen, welche manchmal abgeworfen 1) Vergl. Wiesner, Technische Mikroskopie. 1SG7. p. 111.; Rohstofle. 1873- p. 37G und Die mikroskopisclie Untersuchung (ies Papiers. 1887. p. 28. u. s. w- 2) Höh ne 1 {I.e. 1887. p. 47.) hat zuerst die zweierlei r.astf;iserformen von Brow^^rmetia ■papxjrifera unterschieden. UEBEll WICHTIGE FASEEPFLANZEN JAPANS ETC. 423 wird. Verzweigte Zellenden kommen Läufig vor, die Zweige sind oft mehrmals verästelt. Die Quersclinitte der Faserbündel sind aucli von zweierlei Formen. Die eine bestellt aus grossen, al)gerundeten oder un- regelmässigen Formen ; die anderen sind wenig in Zahl, dick- wandig und polygonal mit abgerundeten Ecken oder sogar abgerundet contouriert. Alle Querschnitte zeigen die aus reiner Cellulose bestehenden Schichten. Die sekundären Verdickungs- schichten sind oft von der primären (i.e. dünnwandigen Scheide) abgetrennt. Das Lumen erscheint breit, abgeplattet oder punkt- förmig. Jodlösung färbt die Bastzellen rothbraun mit dunkler ge- fiirbten Verschiebungslinien, und auf weiterem Zusatz von Schwefelsäure geht diese Farbe in himmelblau über. Die Holz- reagentien geben keine Reaktion. Durch Kupferoxydammoniak werden die Bastzellen sofort in Lösung gebracht. Anhang. Die Bastzellen von Broussonetia papyrifera lassen sich von denen aus B. kasinoki schwer unterscheiden. Ihre Länge beträgt 5.5-11 mm, und die Breite 10-35/^. 9) Edgevjorthia papyrifera. (Fig. 50-52.) Die Bastzellen verbinden sich locker und sind sehr leicht isolierbar. Die Länge beträgt 0.7-4.5 mm, die Breite 14-31 /a Der Umriss der Bastzellen ist höchst variabeP). Eine conti- nuierliche Dickenzunahme von den Enden nach dem Mitteltheil sieht man niemals und fast an jeder Zelle treten plötzliche 1) Dasselbe Vtnliä Unites wurde scliou von K. ISiippi iaii (]]ei träge z. Kenntniss der Tliyme- liaceaî und Poneaecie. Knglcr's Bot. Jalirl». Bd. XVIII. 18i)4. p. oi;!.) und H. Solereder (Systematisclie Anatomie der Dicolyledonen. 1899. p. 812.) bei den Bastzellen von Thymeliaceen beobachtet. 424 K. SAITO : ANATOMISCHE STUDIEN Erweiterungei], Verengerungen und noch Wellungen auf. In allgemeinen a1)Gr haben die meisten Bastzellen überwiegend schmale Enden und breite Mitteltheile, und enthalten in ihren breiten Lumen nur Spuren von körnig auftretenden Plasmamassen. Die Zellendcn sind meist abgerundet, häufig verzweigt, mit dicker Wandung und etwas erweiterten Lumen. Selten kommen spindel- förmige Bastzelleu, welche kurz und breit sind, vor ; bei solchen Zellen erscheint das Lumen schmal und fast gleichmässig weit. Aber an den meisten Bastzellen läuft die äussere Contour der Wandung der inneren nicht parallel. Hierzu tritt aber noch die Eigenthümlichkeit, dass an einzelnen Stellen der Zelle das Lumen ganz verschwindet. Hin und wieder erkennt man kleine Tüpfelspalten. Die Querschnitte sind rund, mit verschieden breitem Lumen. Die Wand zeigt keine Schichtenstruktur. Jodlösung färbt die Bastzellen goldgelb. Auf weiterem Zu- satz von Schwefelsäure bleibt diese Färbung unverändert, aber nur einige werden bläulich gefärbt. Die Holzreagentien geben bei einigen die charakteristische Färbung, während andere nur schwach verändert werden^). Durch Chlorzinkjod konimen zuerst intensiver gefärbte Querstreifungen unregelmässig angeordnet vor. Kupferoxydammoniak färbt die Bastzellen sofort unter starker Aufquellung blau, und zeigen dabei manchmal angeschwollene Partien, die durch Knoten von einander abgetrennt werden. 10) ] Vickslrœmia s ihoh ianum . Die Bastzellen verbinden sich locker ; die Länge beträgt 2.5-5.3 mm, und die Breite 10-30/^. ]Mor[)liologische und ])Bei den Haiulcl.siiKitcik'lii lindct iimn (Ül' Ilolzivaktion der DastzellfU vun Tliyincli acccnarton nicht mehr. UEBEE WICHTIGE FASERPFLANZEN JAPANS ETC. 425 clieraisclie Eigenschaften lassen sich bei den Bastzellen von Wichstrœniia sikohianum nicht deutlich unterscheiden, doch sind sie im allgemeinen dünnwandiger und grosslumiger als bei Edgeioorthia papyrifera, und haben selten Verengerungen. 11) Daphne pseudomezereum. Die morphologischen und chemischen Eigenschaften der Bastzellen bieten kein besonderes Interesse. Die Län";e beträft 1.3-6.2 mm, die Breite 10-25 /A Die primären Bastzellen sind rundlicher und regelmässiger als bei sekundären gestaltet. III. Einiges ueber physikalische und chemische Eigenschaften der Bastzellen. In diesem Kapitel will ich mich mit einigen merkwürdigen Eigenschaften der Bastzellmembranen, i.e. Verschiebungen, Lignin- und Eiweissreaktionen beschäftigen. 1. Ueber die ,, Verschiebungen" der Bastzellen im Sinne v. Höhnel's. An der Wandung der Bastzellen ftmd Reissek^) häufig eigenartige Knoten und Querstreifen, von welchen er die ersteren für eine von umgebenden Elementen hervorgebrachte Bildung hielt und die letzteren für kleine Querspalten und Hohlräume, welche die Verdickungsschichten durchsetzen. Aber die ähnlichen Streifen, die sehr oft an den verarbeiteten Fasern vorkommen, hat 1) Reissek, Die Fasergewebe des Leines, des H;infc8, der Nessel und Baumwolle. (Aus dein IV Bande d. Denkschrift d. Matlieni.-Natiuw. Klasse d. Kais. Akad. d. Wissenschaft. 1S52. p. 12). 426 K. SAITO : ANATOMISCHE STUDIEN er für cine von niecliaiiisclieii Ursaclieii her vor gebrachte Erschei- nung gehalten, Nachher untersuchte VetillarcV) die „Verschie- bungen" im Sinne von Höhnel's (=plis de flexion), doch bleibt nach ihm die wirkliche Ursache noch unaufgeklärt. Höhnel') hat die Querlinien (Streifen) und Knoten der Bastzellwand „ Verschiebungen " genannt und schliesst er, dass dieselbe nichts anderes sind als Bruchstellen, welche von ungleichmässigem Druck des Pllanzengewebes hervorgebracht wurden. Wiesner'^) hielt die Knoten für ,, Kunstprodukt", aber es ist nach ihm noch unent- schieden ob alle, von Höhnel als ,, Verschiebungen " genannten Querlmien, wie der Autor denkt, schon im Pflanzenkörper durch Gewebedruck entstanden sind. Entgegen der Ansicht Höhnel's schrieb Schwendener^) diese Erscheinung nur einer mechanischen Ursache zu, er konnte an gut losgelösten Bastzellen keine Verschiebungen und Kisse finden. Um ein Urtheil darüber zu gewinnen untersuchte ich sowohl die technisch präparierten Bastfasern als auch die Basttheile aus intakten Stammstücke. Zmiächst ist zu bemerken, wie schon von Höhnel gezeigt wurde, dass monocotyle Pflanzen keine Verschie- bung der Bastzellen zeigen, bei den Dicotylen dagegen kommen die Verschiebungen deutlich an den Bastzellen von Urtica ceen {Boehmeria nivecc'\ B. spicala, Urtica Tliunbergkina) , Moracecn 1) Vétillard, Etudes sur les fibres végétales textiles. 1876. 2) Höhnel, Beitrage zur Pllanzeuaiiatomie und Physiologie. Bot. Ztg. Bd. IV. 1882. p. C)21.; Ibd., Ueber den Einlluss des Rindeudruckes auf die Bcschaiienheit der Bastfasern d. Dicotylen. Jahrb. f. wiss. Bot. Bd. XV. 1884. p. 311.; und Ibd., Mikroskoiiic der tetliniscli verwendeten Faserstoffe. 1887. p. 10. 3) Wiesner, Die mikroskopische ITuU-isuchung "). 4) Scliwendene r , Ut-ber die Verschiebung(.n der Bastfasern im Sinnt' \. Jloimel's Ber. d. I). B. G. Bd. XH. 1891. p. 239. 5) Höhnel, I.e. p. 316. UEBER WICHTIGE FASERPFLANZEN JAPANS ETC. 427 {Cannabis sativa^\ Broussonetia kasinoki, B. 'paj)yTifercc^\ Liiia- eeen [TÀnvm usifailsi^îmunf'^), Celastracocn {Celadi'u>^articulalus) vor. All den mittelst Mikrotom hergestellten Längsschnitten der Basttheile war die Beziehung der Verschiebungen zu den umge- benden Zellen klar zu sehen. Diejenigen Theile der Bastzellen, welche Parenchymzellen berühren, besitzen die Verschiebungen genau an der Kontaktstelle der letzteren, und dieselben Bilder, die Höhnel in seiner Arbeit angiebt, wurden auch von mir gefunden. Wie von dem letzt genannten Autor einmal geschah, so un- tersuchte ich auch von entwickehmgsgeschichtlichem Standpunkte aus die Bastzellen. An den jungen, noch dünnwandigen Bastzelle der Dicotylen lässt sich keine Verschiebung nachweisen ; dieselbe kommt in den unteren Theilen des Stammes der krautigen Pflanzen häufig vor, während in dem oberen Zweigen nicht oder nur spärlich. Ganz demselben Unterschied begegneten wir bei jun- gen und alten Zweigen der Bäume, oder bei sekundären und primären Bastzellen. Diese Thatsache erklärt entschieden das nachträgliche Zustandekommen der Erscheinung bei den Bastzellen. Somit kann ich die Angaben von Höhnel völlig bestätigen und es schliesst „ jeden Zweifel darüber aus, dass die Verschiebungen der Bastzellen nichts anderes sind als scharfe Biegungs- oder oft Bruchstellen" (Höhnel, I.e. p. 325), die von dem ungleichmäs- sigen Druck der umgebenden Elemente hervorgebracht wurden. Anderseits aber können diese Verschiebungen, wie Reissek, AViesner und Schwendener untersucht haben, als „Kunst- 1) Höhnel, I.e. p. 316. 2) Ebenda, p. 322. 3) E))cntla. p. 317. 428 K. SAITO : ANATOMISCHE STUDIEN produkt" Iciclit her vorgerufen werden. So fand ich in den im Handle befindlichen oder präparierten Bastfasern von Paeraria Tliunbergiana und Ulmiis montana, var. laciniata die Verschie- bungen (Knoten und Querlinien) in der inneren Celluloseschicht der Wandung. Wenn eine Faser von Flachs, Hanf u. a. künstlich gebogen wird, so kommen unregelmässig hervortretende Knickun- gen mit Querstreifen leicht zu Stande und dies zeigt den Grund warum in verarbeiteten Bastfasern viel häufiger die Verschiebungen vorkommen als bei denselben in le])enden Pflanzen. Daraus schliessen wir, dass die fraglichen Verschiebungen tlieils in Folge von Spannungsverhältnissen schon in den lebenden Pflanzen vorhanden, theils al^er aus mechanischen Ursachen erst beim Präparieren entstanden sein können. 2. Ueber die Ligninreaktion der Bastzellen. Die Wände der Bastzellen bestehen aus gewissen Stoffen, worunter die Cellulose und Holzstoff sehr verbreitet vorkommen. Die Verholzung der Bastzellen kommt durch das frühzeitige Ent- stehen des Holzstoffes in den vorher aus reiner Cellulose zusam- mengesetzten Membranen zu Stande ; doch fehlt sie Ijei gewissen Bastzellarten gänzlich. Oft in derselben Gattung hat eine Art verholzte Bastzellen, die andere aber nicht, z. B. bei Boehmeria nivea sind sie häufig verholzt, dagegen bei B, spicata überall unverholzt. Hinsichtlich der chemischen Eigenschaften des Holzstoffes wurden schon in frülierer Zeit viele Untersuchungen angestellt, doch werden recht abweichende Ansichten von verschiedenen Forschern vertreten. 1) Cza])c k, Ueber die sogenannte Ligninreaktion des Holzes. Sep. Alul. aus Hopj^e- Seyler's Zeitschrift f. pl.ysiol. Ciieniie. JM. XXVH. Heft 1 und 2. 189t>. ri:iîi';i: wiciiTKiK fasi:hpflan/i:x japans etc. 420 Dioso rjiklarlu'it dci- eliemisclipu Natur dv^ verholzton Meiu- l)raiist(tfïes wurde iu aller ueuester Zeit durch die vortreftliehe Uutersuehuui;') Czajiek's heseiti^^'t. Die aus dvn Holzt'eilspäh- ueii uaeh seiner ueueu l)arst:ellun!j,sineth()(le isolierte Suhstauz i;'iel)t die eharakteristisehe Phloroghicinreaktion, und dem chenii- S(dien Verhalten nach ist sie ein aromatiseher Aldehyd, wofür der ^Vutor d(Mi Xamen „Hadromal" voro-esehla^'en hat. .üass die verholzte ^Memltran der I^astzellen das Hadronial enthält, scheint mir ohne Zweifel, inid S(^ suchte ich der C-zapek' sehen Methode f)lgend nach diesem F^toife, und zwai- in allen Fällen mit a;leichem Eid'ol^-e. ])ie Tveaktionsfirhc der extrahierten Li'isuno" weicdit nach dem Verholzuno-sirrade ek, l.r. ISS'.I. 2) Vcr-l C/.ajuk. l.r. 4o0 K. SAITO : ANATOMISCHE STUDIEX 3. Zur Eiweissreaktion der Bastzellmembran. In (1er vorliesfenden Studien hatte ich Gelesienlieit bei einigen, yon mir untersucliten Bnstzellen (Oryza saliva, Bamhusa stenostochia, 3Iiim xapientuiii, var. I'nd-lnen.^ix und Alpinia nutans) die Kothfarbung der Membranen mit JM ill on 's Reagens leielit zu constatieren. Bei anderen monoeotylen und allen dieotylen Bastzellen wurde die Reaktion nieht erhalten. Es ist von Correns') darauf aufmerksam gemacht, ,,dass die Elemente, deren Membranen am stärksten mit Mill on 's Reagens reagieren, mit jenen, die das jNfaxinunn der Verholzung zeigen, gar nicht identisch sind". Auch nach meiner Untersuchung zeigen die nicht oder nur schwach verholzten subepidermalen Bastzellen der Internodien von Oryza sativa eine ebenso starke Reaktion, als dieselben in den Knoten, welche schon stark ver- holzt sind. Aehnliches fand ich l)ei den Rohfasern von Oryza. sativa, welche keine Holzi-eaktion zeigen, aber die Rothlarbung mit ]\Ii lion's Reagens deutlich gel)en. IFalmstucke \o\\ Bamhusa. dciioxtachia, welclie mit Ziniu'hloiiir von dem ITadromal l)efreit waren, gal)eu noch stark die INI illou 'sehe Reaktion. Auch die Reaktionsfarben des Hadromals und der Bastzellmend)ran mit Millon's Reagens w'cichen deutlich von einander ab ; l)ei der ersteren ist sie orangeroth, dagegen bei dem letzteren ziegelroth. Entwickelun2;s2;eschichtlich habe ich die Bastzellen von Bambusa, stenostachia verfolgt. In (k'u weissen Zuwachszonen am Grunde junger Halme tritt in den I^astzellmem brauen mit Millon's Reagens keine deutliehe Färbung ein, während der ])lasmatische Zellinhalt sich intensiv zieo;elroth färbt. Ferner "-iebt die ])C. C'orrons, I'cIkt die ve:?el;>l>ilisclie /cllmcmlirMii. Jnhrl). f. wiss. Rot. T.d. XXVI. UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 431 Meiuhrau in etwas tortgesehrittenen Stadien stark die Mill on 'sehe Reaktion, doeli ist die Verliolzunir lanire nocli nicht heironnen. Diese ïhatsacheii bestätigten völhg Correns'.s Ansieht über die Unabhängigkeit der Reaktion von der Verholznng. Was aber die Ursache der genannten Reaktion anbelangt, so liat Wiesner^) seinerzeit sieh vorgestellt, dass sie wohl anf Vorhandensein des Eiweissstofles bernhe, Avelcher in jenen Zelhvänden enthalten ist, die so lange waehsthumsfähig sind. Dagegen betonen Fi- scher-), Correns'^) und Stras hnrger"") dass die Reaktion hier ineht durch ]^]iweisskörper, sondern durch Tvrosin hervorgebracht wird. Czapek") hielt es wahrscheinlich, dass das von den Mem- branen der Laub- untl Lebermoosen isolierbare ,,Sphagnol" auch die Millon'sclie Reaktion der Zellwände von Bromeliaceen, Zea Mays u. s. w. bedinge. \ov kurzem fand Shil)ata'') im liiesigen Lahoratorium bei den jungen Rastzellt'U v(in lhinihii)^a-\\[v\\ eine reicblicbe Menge von Tvrosinkrvstallen, welche mit der Vei'(Hckung der Baslzellen allmälilich al)nehmen, und sich schliessHch nicht mehr linden lassen. Da er nun nachwies, dass die rothe Reakti(jn der Mem- bran genau diesel! )e Zu- und Abnalime erleidet wie der Gehalt an TyrosinkrystaUen, so ist es höclist walirscheinlicli, dass die Ursache l)Wiesner, Untersuclunigen über die Organization der vegetabilisclien Zellliäute. Silziingsber. d. Kais. Akad. d. Wiss. Bd. XCIII. 1. 1886 ; derselbe, Zur Eiweissreaktion und SUuktur der Zellmembran, Ber. d. D. B. (i. Bd. VI. 1888. p. 33- 2) A . Fisclier, Zur Eiweissreaktion der Zellmembran. Ber. d. D. B. (t. Bd. V. 1887. p. 4"J3.; derselbe, Zur Eiweissreaktion der Membran. Ibid. Bd. VI. 1888. p. 11;;. 3) C 0 r r e n s , I.e. p. 61G. 4) E. Strasburger, Die plhin/.lielien Zellbiiute. .Jabrb. 1". wiss. But. Bd. XXXI. 1898. p. 511. ö) Czapek, Zur Chemie der Zellinemljranen Ijei den Laub- und Lebermoosen. Flora. LXXXVL 1899. Heft. 4. p. 361. 6)K. Shibata, Beiträge zur Waclisthumsgeschichte der Bambus-gewäclise. Abd. a. d. Jonr. of the College of Science, Imperial University, Tokyo, Japan, vol. XIII, pt. III. 1900. p. 483. 4o2 K. SAITO : ANATOMlSCItK SPTDIKN wenigstens bei Bainliusen auf dem A^)rliaii(leiiseiii v<»ii Tyrosiii in der Zellliaut. Damit ist ahcr niclit ausgeschlossen, dass l)ei Bast- zelleii anderer Plianzenarten andere Sul)stanzen diesel) )e Keak- tiun liervurruten. IV. Zur Entwickelimgsgescliichte der Bastzclle Die Uiitersnelnnigen') Ha herla nd t ' s liaben es klargestellt, dass das Bastgewebe sowobl aus jugendliehcr l^^pidermis als aueii aus dem Cambium oder ( Jrundpareneliym liervorgelien kann. Bei den von mir untersueliten Fällen aber traten die Bastzellen stets aus der ( ambiumanlage hervor (Fig. ")•)). 8ie sind zuerst sehr dünnwaiKbg, und werryï/ /<(;/.(/ selir stark, dagegen bei i'rlica Thunlxi-tjinixi und I>aiiiJ)n>astcand>inm/,ellen fand ich niemals (He distinkte stark Hchtbrecln'iide innerste Sc-hicht") 1) II ;i lic !■ I :i 11 il t . Kiil\vii'l<(_'lmii;s;;('scliiclilr (1rs niofliaii. < li'\vrli(s\>toiiis der rilaiizcii. JST'.i. ■J) Sc 11 w «' 11 rlic rriiirip itc- p. •').; 1 F a lir r I ;i ii d i , l.r. \). 00. o) llal)t'r 1 iind 1 , /.c. p. 51. UEr.Kj;, ^\•ICllTl(iK fasehpklanzen jai'anö etc. 4oo der c'ulleiu'hvnuitisclieii NVaiiduiig. Diese Schicht soll lüicli }l;t1)crlaii. iriC. K. SAITO : AXATOMTSCITE STT^PTEX (liuiort SO laiigo fort, als das Lunicii iiocli IcJx'ndcs Plasma fiilirt. Wielitij;' ist die Frai;'e, ol) cine verliolzte Zelle noch in die Tiinge wachsen kann. Xacli den rntcrsnclinn<;'cn von Scliellcn- Lci'g^) besitzt die einmal verholzte Zelle keine Thcilnngsfähigkeit melir nnd eine verholzte Zellmend)ran zeigt kein Fläelieiiwachs- thnm. Xathansohn'") in seinen Untersnehnngen über das Waehs- thnm del' traehealen Kiemente sehliesst aber, dass ,,(ru' \"erhol/nng keine zur Regulierung des Waelisthums (lienen(h' Einrichtung ist ; im (»egentheile , dass das W'aehsthuni regulatoriseh auf die Anlage vei'holztev Elemente einwirkt ". Speeiells in Bezug auf die Bastzelleu aber fehlten es bisher Beobachtungen, welche die Beziehungen zwischeu Verholzung und AVachsthum be- leuchten ; so müssen weitei'c empii'ische Beweise darüber er])ra('lit werden. Bei meinen I'ntersuchungcn beobachtete ich, dass bei rucraria TliunhergUciKi die Länge des Tuternodiums, bei welchen die ver- holzten und uuverholzten Bastzellen gleichzeitig vorkonunen, derjenigen der darunterliegenden gleich war, welche, wie schon oben erwähnt, ausschliesslich Stärkek("»rnei'n versehen waren. l^)ei solchen Tnternodieu wai'cn die Tu] )felgeiä sse schon völlig ausgebildet und die Sti'cckuug des l)eti'eHen(len Internodiums zu dieser Zeit schon vollständig beendet, zugleich l)eganu die Vei'holzung dei- ersten \Van(llamelle der ]>astzellen von unten nacii oben allmählich fortzuschi'eiten ; und so musste die ] ) Schol IciiImm-o-, I.e. 2) A. >' a t li :i nsii li n, ]>oiti-ii,2;e /nr K.iiniiiis-, des Waeli-Iliunw ilcr triU'licalcii ]'".lc'moiif(>. Jiiiirb. r. wiss. i;.,t. i'..i xxxir. isus. p. c.ti. UEBER WICHTIGE FASERPFLANZEN JAPANS ETC. 437 passive Dolinunp; der Zellwand iiacli iliror Yerliolziing ausge- schlossen werden. Was das active Wachsthnm der Bastzellen betrifft, so ist es anifallend, dass die nn verholzten Bastzellen, welche in einem Internodinm mit den verholzten nebeneinander vorkommen, in ihren Enden nocli mit Qnerwänden versehen sind, während bei den verholzten ans demselben Internodium eine durch gleitendes Wachsthum') hervorgerufene, endgültige Aufrichtung resp. ein Steilerwerden der schiefen Endflächen schon vollendet war. Gleiches gilt fin- Corchorus capsularU. Diese Erscheinung, von einem anderen Grunde als demjenigen, welchen Sc h eilen her g angiebt, ausgehend, bestärkte mich in der Meinnng, dass die Bast- zellen nach ihrer Verholzung die Fähigkeit des Eigenwachsthums verlieren. Insoferne nun meine verhältnissmässig wenigen Beobachtun- gen ein Urtheil gestatten, beginnt die Verholzung der Bastzellen dann wenn die Bastzellen ihre passive Dehnung und ihr actives Wachsthum vollendet haben : der Process schreitet dabei von dem nnteren nach dem oberen Theile des Internodiums all- mählich fort, l)is alle Bastelemente verholzt sind. Doch kann ich die Schell enber g 'sehe Ansicht, dass die Verholzung der Wände eine wachsthumshemmende Einrichtung der Zellen ist, nicht gut annehmen, weil die anatomischen Kriterien für Wachs- thumsbefahigung der Bastzellen uns zur Zeit noch unbekannt sind, und für solche Zellen ferner kein wachsthumsregulierendes Mittel mehr nöthig ist ; somit lege ich auf die Verholzung der Bastzellen, Nathans oh n beipflichtend, keinen weiteren Werth als eine mechanische Einrichtung die ihre Aufsteifung herstellt. 1) Vergl. G. Krabbe, Das gleitende Wachsthum bei der Gewebebildnng der Gefäss- pflanzen. 1886. 438 K. SAITO : ANATOMISCHE STUDIEN Obgleich die völlig ausgebildeten Bastzellen meist Plasma nnd Kerne verlieren, so enthalten sie selbstverständlich in jugend- lichen Stadien dieselben stets. Auf die Mehrkernigkeit der Bast- zellen wurde bereits von früheren Forschern aufmerksam gemacht und es soll hier zunächst das Schicksal des Kernes in der Bast- zelle l)esprochen, und dann gezeigt werden, wie die Vermehrung des zuerst einzigen Kernes sich vollzieht. Treul) nahm für die Kernvermehrung der Bastzelle von Urtica dioica eine indirekte Theilung an, während Kallen^) die Vermehrung durch den Fragmentationsprocess vor sich gehen lässt. Bei verschiedenen Bastzellen in jugendlichem Zustande be- gegnete ich vielen Kernen von runden, ovalen oder spindel- förmigen Gestalten ; ferner untersuchte ich in dieser Hinsicht verschiedene Internodien von einem noch unausgewachsenen Stamm von Urtica Thunhergiana, Pueraria Tkunhergiana u. s. w. Um einer Täuschung 1)ei der Beobachtung vorzubeugen, fixierte ich die Objekte mit Flemming'scher Lösung und nach dem Färbender Mikrotomschnitte mit Safranin, Gentianaviolett und Orange wurden die Bastzellen unter dem Mikroskope verfolgt. Die Bastcambiumzelle aus dem Vegetationspunkt trug einen rundlichen Kern, welcher natürlich in diesem Stadium mitotisch sich theilt. Die langgestreckte, aber noch dünnwandige Bastzelle von Urtica Thunhergiana aus älteren Internodien enthält mehrere Kerne, welche fast alle amitotische Theilungsstadien (Fig. 60) und selten die karyokinetische Figur (Fig. 61) zeigten. Interessant ist die Frage, ob diese Karyokinesis nur Vermehrung der Zell- kerne bringt oder gleichzeitig zur Zelltheilung führt. Wegen des seltenen Vorkommens der Erscheinung war es mii- leider nicht 1)F. Kall en, Verhiiltcn des Protoplasmas in dem Gewebe von Urtica urens, entwicke- Inngsgoschiclitlicli dargestellt. Flora. Bd. XL. 1S82. p. 85. UEEEE WICHTIGE FASERPFLANZEN JAPANS ETC. 439 gelungen, eine Entscheidung darüber zugeben, doch halte ich die letztere für das wahrscheinlichere, denn die Bastzellen, welche die Mitosis der Kerne zeigten, sind noch mit Querwänden von einan- der abgetrennt (Fig. 61). Die Formen des Kernes sind man- nigfaltig : kreisrund bis spindelförmig in allen Uebergängen. Die Amitosis kommt nur bei den länglichen Kernen zu Stande, indem sie, wie von K a lien gezeigt wurde, durch Verdünnen von Kern- substanz an einzelnen Stellen, und durch Auseinanderzieheu und endliches Zerreissen zu Tochterkernen werden. In älteren Stadien fand ich noch mehrere kreisrunde oder spindelförmige Kerne, welche aber bei den von mir beobachteten Fällen nie mitotische Theilung zeigten. Die amitotische Kern- theilung fand ich auch bei jungen Bastzellen von Boelimeria nivea, Corchorus capsular is u. s. w., und bei Puer aria Thunhergiana sogar in etwas verholzten Zellen. Die Zahl der vermehrten Kerne nimmt aljer mit der Wand- verdickung der Zelle allmählich ab, und l^ei völlig ausgebildetem Zustande der Zellwandung fand ich sehr oft in Plasmaresten keinen einzigen Zellkern mehr. Die völlig ausgebildeten Bastzellen enthalten, wie schon Hab er 1 an dt zeigte, nur Luft und zuweilen Zellsaft. In einigen Fällen aber z. B. bei den Bastzellen von Boelimeria spicata, lAnum usitatissiiiium eine beträchtliche Menge der Stärkekörner, und bei Hibiscus syriacus kommen Tropfen des Fettes in Bast- zellen vor. Ein sehr häufiger InhaltstofF des Bastcambiums ist aber Ei weiss, welches mit dem Alter der Zelle allmählich abnimmt, und schliesslich verschwindet. Nicht selten enthalten die jungen Bastzellen ausserdem noch Stärkekörnchen, welche Ijei den Bast- 440 K. SAITO : ANATOMISCHE STUDIEN Zeilen von Pueraria Thunhergiana. nach ihrer Verholzung noch unverändert l)leiben. Von anorganischen Stoffen fand ich in den jungen Bastzellen stets Magnesia und Phosphor säure'), welche in den völlig ausgewachsenen Zellen nicht mehr sich nachAveisen lassen. Ein Abnehmen luid eventuelles Verschwinden dieser Stoffe findet auch nach etwaiger Verholzung statt. V. Uebersicht ueber die präparierten, in den Handel kommenden Bastfasern. Die mikroskopischen Untersuchungen der europäischen Han- delsfasern verdanken wir in erster Stelle Wiesner'-). In neuester Zeit erschien der vortreffliche Katalog von Dodge^), welcher aber der Natur seines Werkes gemäss sich hauptsächlich mit der Kultur, Verberei tung u. s. w. beschäftigte und nur wenig die Histologie der Pflanzenfasern behandelte. In Folgendem gebe ich eine Uebersicht über die Neben- bestandtheile, welche in den japanischen Pflanzenrohfiiseru mit den Bastzellen gleichzeitig vorkommen, an, hauptsächlich vom histologischen Stand})unkte ; die Art und Weise der Verarbeitung der Fasern u. s. w. ist natürlich umberührt gelassen. 1) Zum Nadiweise von ^lagiiesia, l'liusp hur«äii re uiul aiulorcii anurgauisoheii Stoflen hedicnte ich niicli der von Zimmciiiia nn {Ix.) angegebenen ]\[cth(Hle. 2) Wies ne r, Ilolistotl'e. 1873. zweite Aufhige ist im Erscheinen begriffen, o) Dodge, Ix. UEBER WICHTIGE FASEKPELANZEN JAPANS ETC. 441 1. GESPINNST- UND SEILV/AAREN. a. MONOCOTYLEDONE FaSER. (MiT GeFaSSE.) TVT - T.ii Beschatleiiheit >. , , , , ., JNaiiic (1er Pflanzen. , r> . n i Aebenbestandtheilo. der Biistzellwand. Oryza saliva. unverliolzt. Spiral ge fasse, Parenchym und verkieselte Epidermiszellen 1). ^Fusa sapienium, var. liukiu- ensis. verholzt. Stärke oder Calciunioxalat führende Parenchynizellen, Stegniata'^), und selten Spiralgefässe, Epiderniiszellen und M ilciigefdsse. Agave amerieana. » Spiralgefässe, Parenchynizellen. Alpinia nutans. » Spiralgefässe, Stärke führende Paren- chynizellen. Pandaims uduraiissvmuLs. V Spiralgefässe, Parenchyinzellen mit ver- holzten und getüpfelten AVänden, stärkeführende dünnwandige Paren- chyinzellen. h. Dicotyledone Faser. (Ohne Gefässe.) Name der Pflanzen. Beschaffenheit der Bastzellwand. Neben bestand 1 1 1 eile. Linum usitatissimum. unverholzt. I'arenchymzellen und selten P]pider- miszellen. Cannabis saliva. halb verholzt. Calciumoxalat führende Parenchyni- zellen und selten Epiderniiszellen. JBoehmcria nivea. unverholzt. Calciumoxalat führende Parencliym- zelleii. B. spicata. )) » Urtica Thunhergiana. )) )) Corchorus capsularis. verholzt. Calciumoxalat führende Bastparenchym- zellen und spärliche Markstrahïen- zellen. 1) Wiesner, Technische Mikroskopie. 1867. p. 2,35. 2) Ueber das Vorkommen der Stegmata bei den il/«.srt-faseni vergleicli Wiesner, Rohstoffe 1873. p. 434. und Hölmel, I.e. 1887. p. 50. 442 K. SAITO : ANATOMISCHE STUDIEN Name der Pflanzen. BescliMlTènlieit der Bastzellwand. Nebenbestandtlieile. Abutilon Avicennœ. verholzt. Calciumoxalat führende Parenchymzellen. Hibiscus syriacus. )) » Puei-uria Tlmnhcrcjiana. liallj verbolzt. Calciumoxalat führende Parenchynizellen und verholzte Sklerenchymzellen^). Wistaria chùiensis. » Zweierlei Arten der Parenchymzellen ; a) kleinzellig aber dickwandi^ç und Calciuiuoxalat führend, 6) grosszellig, aber dünnwandig und stärkefülirend. Auch selten verholzte Zellen. (Ilnms viontana, var. lacini- ata. )> Calciumoxalat fülirende Parcuchymzellen. Tilia cordala, viiv. Japonica. verliolzl. Getüpfelte und Calciumoxalat führende Bastparenchymzellen ; getüpfelte und stärkefülirende Markstrahlenzellen, und langestreckte Parenchymzellen mit getüpfelten und verholzten Wänden. Finninia platanifolia. )! Calciumoxalat oder Stärke führende Parenchymzellen und Stein zellen. Vilis Coigndiir. » Stärke fülirende Bastparenchymzellen und Stärke oder Calciumoxalat führende Markst rahlcn/.ellen. 1) Diese Sklercnchymzellen gelten zur guten Erkennungsmerkuiale der Faser von Pueraria Tliunbcnjiana. UEBER WICHTIGE FASERPFLANZEN JAPANS ETC. 443 2. PAPIER. , ^, Beschaffenheit î^arae der Pflanzen. j t> . n j der Bastzellwand. Nehenbestandtheile. Brows> !) Oryza sativa. » Verkieselte Epidermiszellen und Bruch- stücke der Gefässe. J3ambusi slenoslaehia. verholzt. Zweierlei Parenchymzellen ; o) dick- wandig und getüpfelte, b) dünn- wandige. Auch Poren- und Netz- gefässen. 3Iasa saplenfum, var. liukiu- e7lMS. )! Stegmata, Spiralgefiisse, Milchgefässe und Epidermiszellen. VI. Résumé. 1. Betreffs der Verbreitung und Anordnung der Bastzellen wurden einige weitere Beiträge zur Vervollständigung früherer Angal)en erbracht. 2. Dimensionsverhältnisse der Bastzellen : — 444 K. f?AITO : ANATOMISCHE STUDIEN Läi ge Breite Name of Pflanzen. m. m. ÎJ.=ra. m. m. Minimum. Maxinuim. Minimum. Maximum. Pandanus odoratissimus. 0.75 2.15 15 25 Oryza /tativa. 0.55 1.90 4 15 Bambusa stenostachia. 0.70 2.80 7 25 A(/ave americana. 0.70 1.90 20 40 3Iusa sapienlum, var. Uuhiuensii<. 2.65 6.40 18 31 Alpinia nulaif^. 0.60 2.70 10 25 Ulmus montana, var. lachilata. 1.50 7.50 10 20 Broussonelia kasinnki. 1.51 10.00 10 ;u B. fapyrlja'a. 5.50 11.00 10 35 Cannalis nativa. 7.00 50.00 10 35 Bo'hmeria nivea. 12.30 245.00 40 90 B. spicaia. 7.00 26.00 11 72 Urtica Thunb&'giana. 5.00 60.00 20 03 Pueraria Thunbergiana. 0.95 4.20 10 22 Wistaria chinensis. 1.30 3.70 10 20 Linum usitatissimum. 14.00 85.00 18 25 Celasfrus articulatus. 20.00 70.00 80 135 . / primäre Bastzelle. Vilis C(ngnetiœ\„ , ,.. „ , ,, ^ »Sekundäre Bastzelle. 1.00 3.00 25 30 0.40 0.95 10 25 Corchorua capsular i^. 0.60 6.35 13 22 Tilia cordata, var. japonica. 1.48 2.40 17 23 Abutilon AvicenncE. 1.00 2.10 8 37 Urena lobala. 0.75 2.43 15 26 Hibiscus syriacus. 0.60 1.70 12 35 Firminia platatiifoUa. 1.50 3.00 15 20 Daphne pseudotnezereum. l.;]0 0.20 10 25 Edgeworthia papyrifera. 0.70 4.50 14 31 Wickslrœmia sikokianum. 2.50 5.30 10 30 3. liniiion mit Vcrongerimgen kommt bei don Bastzellen von Boehmcria npicata, Corchorus capsularis, AbutUon Avîcennœ, Urena lohata, Hibiscus syriacus, Firminia platanifolia, Edgeworlhia 'pa2oyrifera, Wiclcdrœmia silcoJcianum und Daphne p^eudomczereum vor ; Lmnen mit Erweiterungen l)ei den Bastzellen von Linum usitatissiviwN und Boehmeria spicata. An den local erweiterten Stellen der Bastzellen von Boehmeria spicata und Liinim imtatissimuni wii'd die Wand UEBER WICHTTOE FASERPELANZEN JAPANS ETC. 44") bedeutend dünner und die Plasmapartien in diesen Erweiterungen pflegen friUier oder später sich einzukapseln. 4. Ein oder mehrere Querwände kommen l)ei den Bastzellen von Jltix Coignetia\ Wistaria chinensis, Oryza saiiva, Banihusa sienostachia, Musa sapientum, var. liukiucnsis, Fandamis odoratis- shniis vor. Die die Wand durchsetzenden Poren sind verschieden gerichtet (linkssehief oder längslaufend), und mannigfach gestaltet (rund oder sj)altenförmig). '"). Die „Verschiebungen" der Bastzellwand sind sowohl in den lebenden Pflanzen durch den ungleichmässigen Druck als auch in präparierten Handelsmaterialien vorhanden. Dieselben fehlen bei allen untersuchten monocotylen Faserpflanzen und auch ]w\ vielen dicotylen Gespinnstpflanzen. 6. Die Verholzung der Bastzellwaud fehlt bei Boehneria spirala, Urtica Thunhergiana, Broussonetia kasinoki, Celastrus arti- culatns und Linum iisitatissiimcm. Ferner die Verholzung der Bast- zellen von Pueraria Thunhergiana, Wistaria chinensis, Cannabis sativa und Ulmiis montana, var. laciniata beschränkt sich auf die äussere Wandlamelle. Die Bastzellen von Vitis Coignetiae und Tilia cor data, var. japonica ßirben sich auf dem Querschnitte der Stengel mit Salz- säure (ohne Zusatz von Phloroglucin) roth. 7. Nach der Czapek 'sehen Methode konnte das Hadromal aus allen verholzten Bastzellen extrahiert werden. 8. Die Mil Ion 'sehe Beaktion der Zellwand wurde nur bei den Bastzellen von Bamhusa stenostachia, Oryza sativa, 3Iusa sapienturii, var. liuhiuensis und Alpinia nutans constatiert. In diesem Falle ist die Färbung von der Verholzung unabhängig. 9. Die meisten der völlig ausgebildeten Bastzellen enthalten Luft und zuweilen noch etwas Plasmareste. In einigen Fällen 446 K. PA TTC : ANATOriRCTTE RTUDTEX nbor f'lilirt dio Bni^tzolle doeli noch StärkekünicT {Linum usiia- tissl/jniJ/f, Borhiiieria f^picata), Fett {Hihiscv^ sijriacus^) und sogar Zellkorno in ihveni Plasmakörper {Alpinia nutans u.a.). 10. Die jungen Bastzellen haben eine eoUenelivm a tische Ver- dickung in ihren Ecken und sind plasmahaltig mit einem oder mehreren Kernen. 11. Die Vermehruno; der Kerne »'cschieht durch direkte Thei- lunir, al)er in den iünsieren Stadien fand ich noch deutliche karv- okinetische Theilung. 12. Das Dicken wachsthum der Bastzellen kommt dadurch zu Stande, dass die neuen Lamellen an der inneren Seite der alten Wand — durch Apposition — angelagert werden. lo. Die Verholzung tritt in der Bastcambiumzelle dann ein, wenn die letztere noch dünnwandig ist, und ihre Enden aber schon völlig aufgerichtet sind, und Plasma noch vorhanden ist. 14. Die jnngen Bastzellen enthalten Eiweiss, Magnesia, Phosphorsäure und zuweilen Stärkekc'irnchen. Anhang. Tabelle zur BESTT:\ri\ruNG vox .tapaxlschen Pflaxzenfaseen. Die analytische Bestimmungstabelle der europäischen Spinn- stoffe wurde von Schlesinger^), Vetillard'), Höhnel") und Behrens*) in den Dienste des technischen Zweckes gestellt. Das Bedürfniss, durch die histologische Methode die japanischen, im 1) Seh Icsi n gor, Examen niicrosoopiqiu' et inicrocliimi(iiU' dos fibre.-; textiles. ISTö. 2) Yr t 1 1 1 n 1(1, Etudes sur les fil)res vésotiilos textiles. ISTC». 3)IIühnel, Mikroskopie der teeliiiiseli verwendeten Faserstollé. 1S87. 4)n. lîchrens, Mikroolicmiselie Analyse organisclier Verliindnni^en. zweite Anllage. 1S9G. Er hat dieses VerfMJii'cn ilincli iiinra-seiide .Vnwcndinit; i)liysilser zeigt selten Gefässe ! Die Verholzung tritt bei einigen Arten nur an den ilusseren Wandschichten.) I. — Nur an den clusseren Wandschichten verholzt. (Cannabis satiua, Uhiius montata, var. laciniata, Fuerarla Thiinbci-giana, JVistaria chinensis.) 1. Querschnitte. Immer in Gruppen angeordnet, mit mehr oder minder abgerundeten Ecken, schUessen dicht an einander. Alle sind von einer dünnen verholzten Aussenschicht umgeben. Das Lumen linienförmig, einfach oder verzweigt, unregelmässig, manchmal mit einspringenden Win- keln, ohne Inhalt, Schöne concentrische Schichtung. Längsansicht. Fasern unregelmässig dick, oft mit daran hängenden Stückchen der Umrandungslamelle. Verschiebungen häufig. Streifungen d(.'ut]ieh. Das Lumen ist schmal oder breit, meistens grösser als bei Linnm nsitat'iashnum. Die Enden sind breit, dickwandig und abgerund<'t, häuHg verzweigt. Länge 7-50 mm ; Breite 10-35//.. Cannabis sativa (Nom. ja[). Asa). (Fig. 53-56.) 2. (..Mierschnitte. Polygonal, geradhnig begrenzt. Sie sind von einer dünnen verholzten Mittellamelle umgeben. Das Lumen })unkt- oder linicnf^Jrmig, selten etwas verbreitert. Nie concentrische Schichtung. Längsansicht. Fasern schmal, oft mit daran hängenden Stückchen der Umrandungslamelle. Verschiebungen häufig an der inneren Schicht, von welcher die äussere Schicht oft abgetrennt und spiralig gestreift. Lumen linienförmig. Die Enden sind abgerundet. Länge 1.5-7.5 mm ; Breite 10-20 n. Uhaus monUma, var. laciniata (Nom. jap. Ohio). (Fig. 31 und 32.) 3. Querschnitte. Polygonal, geradlinig begrenzt, selten etwas ab- gerundet, von einer relativ dickeren Mittellamelle umgeben. Lumen [)unkt- förmig oder breit. Längsansicht. Fasern schmal, mit daran h;üigenden Stückelu'n der Mittellamelle. Verschiebungen liäulig an der innc.'ren Schicht. Lumen linienförmig oder l)reiter, mit Plasmaresten. Enden sind stumpf oder etwas abgerundet, manchmal verzweigt. Länge 0.95-4.20 nnn ; Breite 10-22 //. TTEP.EK WTCHTTGE FASEPtPELAXZEN JAPANS ETC. 4'")1 (An den Handelsraaterialien liommen stets grosse getüpfelte Skleren- ch^-nizcllen vor.) Pneraria Thiinhcrgiana (Nom. jap. Kiulii). (Fig. lG-18.) 4. Quersclmi tte. Polygonal, geradlinig begrenzt, von einer dünnen Mittellanielle umgeben. Lnmen breit und rnnd. Längsansicht. Fasern schmal, oft mit daran hängenden Stückchen der Mittellarnelle. Verschiebungen hantig an der inneren Schicht. Lnmen meist lireit, oft mit Plasmaresten. Die Enden hänfig in innere und äussere Schichten getrennt, stumpf, manchmal verzweigt. Länge 1.3-3.7 mm ; Breite 10-20//. (An den Handelsmaterialieu von rr/s^or/a-faserri kommen stets in lleihen angeordnete, Calcinmoxalatkrystalle einschliessende Zellen vor). Wistaria chincnsis (Nom. jap. Fuji). (Fig. 35 und 3G.) II, — Granz verholzt. a. Lumen mit auffiillenden Verengerungen. {Corchoriffi capsularis, Ahntilon Avicenno', Hibiscus syriacus, IJrcna loJtnta, Firminia plaianifolia.) 1. Querschnitte. Gruppenweise angeordnet, polygonal, geradlinig begrenzt ; Ecken scharf. Lumen rund oder oval, glatt, ohne Inhalt, Längsansicht. Fasern glatt, ohne Verschiebungen und Streifungen ; Lumen deutlich sichtliar, breit, mit Verengerungen, verschwindet aber nie. Die Enden immer abgerundet und massig stark verdickt, weitlumig. Länge 0.0-6.35 mm ; Breite 13-22 //. Corcliorus capsularis (Nom. jap. Tsunaso). (Fig. 4.3 und 46.) 2. Querschnitte. Im allgemeinen etwas grösser als bei Corcliorus capsularis, geradlinig begrenzt, oder etwas abgerundet. Lumen rund oder oval, grösser als bei Corcliorns capsularis. Längsansicht. Fasern ungleichmässig dick, glatt, ohne Verschie- bungen und Streifungen. Lumen gross, mit auffallenden Verengerungen, und stellenweise ganz fehlend. Enden stumpf, stark verdickt, häufig ver- zweigt. Länge 1-2.1 mm ; Breite 8-37 //. Ah/dilon Avicowo' (Nom. jap. Ichibi). (Fig. 29 und 30.) 3. Querschnitte. Polygonal, geradlinig begrenzt, mit abgerundeten locken. Lumen rujid oder oval, selten etwas eckig. 452 K. SAITO : ANATOMISCHE STUDIEN Längsansicht. Die Breite an einer und derselben Faser sehr un- gleich. Lumen sehr breit, selten mit Verengerungen. Enden stumpf oder verzweigt und nicht weitlumig. Die Wand ist im allgemeinen dünn. Länge O.G-1.7mm; Breite 12-35 /^ Hibiscus syriacus (Nom, jap. Mukuge). (Fig. 25 und 2G.) 4. Querschnitte. Mehr oder minder polygonal, mit scharfen oder abgerundeten Ecken. Lumen klein, punktförmig, zuweilen breiter und oval. ]\Iittellamelle breit. Längsansicht. Die Dicke ist an einer und derselben Faser un- gleichmässig. Lumen meist schmal, mit auffallenden Verengerungen, stellen- weise ganz fehlend. Enden stumpf und stets verdickt, aber nicht autfanend weitlumig, Ku[)feroxydammoniak bewirkt fast gar keine Auhpiellung. Länge 0.75-2.43 mm ; Breite 15-2G //. Urena lohata (Nom. jap. Obondenkwa.) (Fig. 33 und 34.) 5. Querschnitte. Polygonal, geradling begrenzt, kommen in Gruppen. Ecken scharf oder etwas abgerundet. Lumen sehr schmal, punktförmig oder etwas erweitert. Längsansicht. Dicke an einer und derselben Faser ziemlich gleich- massig. Wandung ist dick und mit runden Porenkanälen. Lumen meist linienförmig und stellenweise ganz fehlend, selten mit mittlerer angesch- wollener Partie. Die Enden stumpf, verdickt, und häufig mit Verzweigungen. Länge 1.5-3 mm ; Breite 15-20 //. Firminia platanifolia (Nom. jap. Aogiri), (Fig. 27 und 2S.) ß. Lumen ohne Verengerungen. {Tilia cordata, var, japonica, Vit is Coiastzelle), 0.4-0.9.) mm (Sekundäre Bastzelle) ; lîreite 2.5-30 // (Primfu-e P,aRtzelle), 10-2.5 // (sekun- däre Bastzelle). Vil/'s Co/gnct/'œ (Nom. ja]). Yamabudo). (Fig. 47-40.) /;. ]\IONOCOTYLEDONE BASTFASERN. (Xeben d(Mi Bastzellen kommen stets Gefässe vor, mit einer Ausnahme von Jf/fsa-fa^er.) I. ]\Iillon'sche Reaktion positiv. {ßamhufid stcnoHtdclila, Musa sapientum, var. Ji/fl'/xenfu's, Alpinîa nutans.) 1. Querschnitte. Meist abgerundet; Lumen immer rundlich und breit oder sehr schmal bis punktförmig. Zweierlei Arten der Bastzellen, dünne und dicke, theils klein, theils gross und weitlumig. Längsansicht. Zweierlei Bastzellen, dünn- und dickwandige. Gleichmässig breit, glatt, mit kleinen Porenkanälen. Enden stumpf. Lumen gross und häufig mit Querwand, Länge 0.7-2.8 mm; Breite 7-25 /y.. Baonhf/sa stenostachia (Nom. jap. 8hichiku). (Fig. 2-4.) 2. Querschnitte. Polygonal mit abgerundeten Ecken, schliessen meist dicht aneinander. Lumen gross, fast oder ganz rund. Längsansicht, Fasern gleichmässig dick, glatt, dünnwandig. Lumen gross und nach beiden Enden allmählich verschmälert, selten mit Querwand. Enden scharf und verdickt. Länge 2.6.5-6.4 mm ; Breite 18-31 //. (An den Bohfasern und Papieren aus dfusa sapientum, var. Jhil-hi- enfiis findet man stets tStegmata, welche in der Asche leicht nachzuweisen sind.) Musn sap/cnfi/ni, var. Ii//l-iacnsfs (Nom. jap. Ito-basio.) (Fig. 14 und 1.5). 3. Querschnitte. Polygonal, geradlinig begrenzt, manchmal mit ab- gerundeten Ecken, dicht aneinander schliessend. Lumen gross, meistens rund. 454 K. SAITO : ANATOMISCHE STUDIEN L an g S a n S i cil t . Fasern gleichmilssig didi, manchmal wellig contonriert, mit Lilngsspalten, Lumen gross, rnnd. Enden l)reit nnd verdickt. Länge O.G-2.7 mm; Breite 10-25//. À/pin/rt nutans (Nom. jap. G(oto). (Fig. 11-13). II. — ]\[ i 1 1 0 n ' sehe Pteation negativ. {^Pandanu.s odoratisftirnvs, Ar/ave amcrlvana.) 1. Querschnitte. Polygonal, geradlinig hegrenzt, mit scharfen Ecken, diclit aneinander schliessend. Lumen rund oder oval. Längsansicht. Mitteltheil einer und derselben Faser auffiillend hreiter. Lumen ziemlich breit. Enden lireit nnd verdickt. Wand mit linksschiefen Porenspalten und manchmal wellig contouriert. Kupferoxydammoniak bringt sie nicht in Aufquellung. Länge 0.75-2.15 mm; Breite 15-25 /i. (An den Eohfasern kommen stets getüpfelte Parenchymzellen vor.) Fand onus odorat iss-im f/s (^om. Jap. Adan). (Fig. 7 und 8.) 2. Querschnitte. Polygonal, geradling begrenzt, scharf eckig, dicht aneinander schliessend. Lumen gross, oval, mit etwas scharfen Ecken. Liingsansicht. Die Breite der Fasernnach den Mitteltheil auffallend gWisser, manchmal wellig contouriert. Limien breit. Enden stumpf nnd verdickt. Wand mit linksschiefen Porenspalten. Kupferoxydammoniak bringt si(^ zu geringer Aufquellung. Länge 0.7-1.9 mm ; Breite 20-40 //. Agave amerieana (Nom. jap. Kiuzetsuran). (Fig. 9 und 10.) Die vorstellende Arbeit wurde auf VeranLnssung nnd unter Leitung des Herrn Prof. Dr. IMiyoslii in einer ZeitiVist von Sept.em]>er 1890 l)is Juni 1000 im l)()tanisehen Institut der Kaiser- li(dien Universität zu Tokio ausgefülirt. Es ist mir eine angenehme Pflicdit, meinem hochverehrten Lelirer für die freundliche Anre- gung und ITnterstützung meinen verbindlielisten Dank auszuspre- chen. Auch Herrn Prof. Dr. Matsumura spreche ich für seine vielfache Belehrung hier meinen herzlichsten Dank aus. UEliEll WICUTRJE i AttEilPELANZEN JAPANS ETC. 455 HeiTii Prof. Dr. Miyabe und Herrn S. Huzuki in Sapporo, Herrn K. Ando und Herrn H. Kuroiwa in Liukiu, Herrn Y. Tanaka, Herrn Y. Shirasawa und Herrn J. Unie mur a, welche mir Alkohol- und Handelsmaterialien auf freundlichste Weise zusandten, l)in ich ebenfalls zu grossem Dank verpflichtet. Botanisches Institut Kaiserl. Uni\-ersität zu Tokio. Decemljer 1900. Litteratur Verzeichniss. 1852. )S. Keissek, Die Fasergewebe des Leincï^, des Hanfes, der x^e,s.sel uutl Baumwolle. Denkschrift d. AVien. Akad. 1853. IL Scliaclit. Die Prüfung der im Handel vurkomnienden Gewebe durch das Mikrosko[) und durch cliemische Eeagentien. 18G5. J. Wiesner, Älikroskopische Untersuchungen der Maislische und der Maisfaserprodukte. Besonderer Abdruck aus Dingler 's polytechn. Journal, erstes Februarheft, Bd. CLXXV. ^. 225. 18G7. , Technische Mikroskoj)ie. 1870. , Ijeiträge zur Kenntniss der indischen Faserptianzen, etc. Sitzungsberichte d. Kais. Akad. d. Wiss. Bd. LXII. 1. 1873. , Die Rohstoffe des Pflanzenreiches, Erste Auflage. Leipzig. 1874. h). Schwenden er. Das mechanische Princip im anatomischen Bau der Monocotylen. Leipzig. 1875. Pi. Schlesinger, Examen microscopifjue et microchimique des fl1)res textiles. l*aris. 1876. M. Vé til lard. Etudes sur les fibres végétales textiles. Paris. (Skiz- zirt von W. H. Seaman in citierten Werke Dodge's, p. 352. Appen- dix B.) 456 K. SAITO : ANATOMISCHE STUDIEN 1877. De Bary, Vergleichende Anatomie der A^egetatiousorgane der rhaneroganien und Farne. Lei[)//ig. 1879. G. Haberlandt, Eutwickelungsgeschiclite des mccluinisdiun Ge- wubesj^stems der Pflanze. Leipzig. 1882. F. K allen, A^erhalteu des Prütoi)]asmas in dem Gewebe von Urtica urens, entwickelungsgescbichtlicli dargestellt. Flora. ]jd. XL, j). G.5. . F. Höhnel, Beitrage zur Ptianzenanatomie und Pliysiulogie. Bot. Ztg. Bd. IV. p. 621. 1883. V. Berthold, Ueber die mikroslvopischen Merkmale der wichtigsten Pflanzenfasern. Beilage zur Zeitschr. für landw. Gewerbe. (Picf. in Just's Bot. Jahresberichte. 1883.) 1884. F. Höhnel, Ueber den Einflüss des Pündeudruckes auf die Beschaf- fenheit der Bastfasern der Dicotylen. Jahrb. f. wiss. Bot. Bd. XV. p. 311. 1885. A. Tscliirch, Beiträge zur Kenntniss des mechanischen Ge- wehesystems der Pflanzen. Ebenda. Bd. XVI. p. 303. .1886'. II. Focke, Mikroskopische Untersuchungen der bekannteren Ge- spinnstfasern, der Shoddywofle und des Papiers. Archiv der Pharmacie. p. 609. . J. Wies nor, Untersuchungen liber die Organisation der vegetabilischen Zellhaut. Sitzungsberichte d. Kais. Akad. d. Wiss. Bd. XCIII. 1. . G. Krabbe, Das gleitende AVachsthum bei der Gewebebildung der Gefäss})flanzen. Berlin. 1887. F. IL'thnel, Mikroskopie der technisch verwendeten Faserstofle. ^^'ien, Pest, Leipzig. . J. Wiesner, Die mila-oskoi/ische Untersuchung des I'a[»iers, mit besonderer Berücksichtigung der ältesten orientalischen und europäischen Papiere. Wien. . G. Krabbe, Ein Beitrage zur Kenntnis« der Struktur und des Waclis- thums der vegetabilischen Zellhäute, Jahrb. f. wiss. Bot. Bd. XVIII. p. 346. . A. Fisclier, Zur Eiweissreaktion der Zellmembran. Jîerichte d. D. J'.. G. Bd. V. p. 423. J888. , Zur EiweissreaiUion der i\leml>ran. ]'vl)enda. Ld. \'l. p. 113. . d. Wiesner, Zur Eiweissreaklion und Stiiiklur der Zelhiicmliraii. Ebenda Bd. \\. p. :'.;]. 1889, A. Tscliirch, Angewandte i'Üan/enanatumie, Wien und Jjei[)zig, UEBER WICHTIGE FASERPELANZEN JAPANS ETC. 457 1S91. Th. Lange, Beiträge zur Keuntiiiss der Entwlckeluug der Geftisse und Tracheiden. Flora. r>d. XLIX. jk 303. 1892. A. Zimmermann, Die botani.^che Mikrotechnik. Tübingen. 1894. C. Correns, Ueber die vegetabilisclie Zellmembran. Jahrb. f. wiss. Bot. Bd. XXA'I. p. 587. . S. Schwendener, Ueber die ,, Verschiebungen " der Bastzellen im Sinne v. Höhnel's. Berichte d. D. B. G. Bd. XII. p. 239. . K. Supprian, Beiträge z. Kenntniss der Thymeliaceai und Peneaceiv3. Engler 's Bot. Jahrbücher. Bd. XVIII. p. 313. . A. Kosoll, Ueber vegetabilisclie Faserstoffe. Jahresberichte der niederösterreichischen Landoberrealschule etc. in Wiener Neustadt. (lief, in Bot. Centralblatt. Bd. LX. p. 215.) 1S96. H. Behrens, Mikrochemische Analj'se organischer Verbindungen, Heft II. . G. Haberlandt, Physiologische Pllan/.enanatomie. Zweite Aullage. . H. Schellenberu', Beiträü;e zur Kenntiiiss der verholzten Zell- membranen. Jahrb. f. wiss. Bot. Bd. XXIX. p. 237. . W. Futterer, Beiträge z. Anatomie und Entwickehmgsgeschichte der Zingiberacere. Bot. Centralblatt. Bd. LXVIII. p. 2-4:1. 1897. C. K. Dodge, A descriptive catalogue of useful liber plants of the world, including the structural and economic classiHcation of fibers. lleports. no. 9. U. S. department of Agriculture, fiber investigation. 1895. W. Behrens, Tabellen zum Gel)rauch bei mikroskopischen Arbeiten. Dritte Aufiage. Braunschweig. J899. A. Na than söhn, Beiträge z. Kenntniss des Wachsthums der trachea- len Elemente. Jahrl). f. wiss. Bot. Bd. XXXII. p. (J7I. . E. Strasburg er, Die pflanzlichen Zellhäute. El)enda. Bd. XXXII. . II. Soler eder, Systematische Anatomie d. Dicotyledonen. . F. Czapek, Ueber die sogenannte Ligninreaktion des Holzes. Separat- Abdruck aus Hoppe-Seyler's Zeitschrift f. physiolugische Chemie. Bd. XXVII. Heft 1 und 2. . , Zur Chemie der Zellmembranen bei den Laub- und Lebermoosen. Flora. Bd. LXXXVI. Heft. 4. p. 361. 190U. K. Sil ibata , Beiträge zur Wachsthumsgeschichte der Bambusgewäehse. Abdruck aus dem Journal of the College of Science, Imperial University, Tokyo, Ja[)an. Vol. XIII. pt. III. 458 K. SAIÏO : ANATOMISCH STUDIEN Inhalt. Seile. 1. Einleitung 395 II. Die Anordnung der Bastzcllen mit ihren histulogischen Merkmalen 401 IIL Einiges über i)liyöilva]isclie imd cliemisclie Eigenscliaften der Bastzellen 425 IV. Zur Entwickelungsgeschichte der Bastzelle 432 V. Uebersicbt iUjer die i)rä[)arierten, in den Handel kom- menden Bastfasern 44U VI. Itésumé 443 Anbang. Tabelle zur Bestinmuujg von ja])aniscben Ptlanzeniasern 44() Litteratur Verzeicbniss 455 ERKLÄRUNG DER TAFELN. TafeL XX. Fig. 1. (x2.")): (^)iier.sclinitt durch den Halm von (Irt/za mtlva. Fig. 2-4. Die Bastzelle von Bamhn^a denostacliia. Fig. '1. ( X 39.3) : Längsansiclit des Mitteltheiles. a dick-, h dünnwandige Zelle, p Porentüpfeln. Fig. 3. (x39.")): Zellende, a, h, p wie oben. Fig. 4. ( X 39.3) ; Querschnitte, o, h wie oben. Fig. 5 nnd G. Die Bastzelle von Orij-in f^athn. Fig. 5. ( X 39.5) : a Längsansicht des Mitteltheiles. h Zellende, p Poren- spalten. Fig. 6. ( X 39:3) : Querschnitte. Die Zelle bei h etwas grösser als bei a. Fig. 7 nnd 8. Die Bastzelle von Pamhinus odoratlssirims. Fig. 7. (x395): Querschnitte. Fig. S. ( X 39.3) : a Längsansicht des Blitteltheiles. h Zellende, j) Poren- spalten. Fig. 9 und 10. Die Bastzelle von Agavo americnno. Fig. 9. (x39.3): Quersclinitte. Fig. 10. ( X 39.3) : a Längsansicht des Mitteltheiles. h Zellende. Fig. 11-13. Die Bastzelle von Alpinla nutans. Fig. 11. (x39.3): Längsansicht des Mitteltheiles. a Zelle mit Längsspalten (p). h Zelle mit Kern (/i-). Fig. 12. ( X 39.3) : Zellende, a normal, h wellig contouriert. Fig. 13. ( X 395) : Querschnitte, a dünn-, h dickwandige zelle. Fig. 14 und 15. Die Bastzelle von Musa sapientum, var. liukmensis. Fig. 14. ( X 395) : Querschnitte. Die Zelle bei b grösser als bei a. Fig. 15. ( X 39.3) : «Längsansicht des Mitteltheiles. ?; Zellende, c Stegmata. p. Poren spalten. Fig. IG-IS. Die Bastzelle von Pueraria Tliunherglana. Fig. IG. (x395): Längsansicht des Mitteltheiles. Lumen bei a grösser als bei h. an äussere, in innere Schicht der Wand. Fig. 17. ( X 395) : Zellende, a ^normal, h mit Verzweigung (?j). an, in wie oben. Fig. 18. (x39.5): Querschnitte, an, in wie oben. Fig. 10 lind 20. Die Bastzelle von Linum nsifatissiomim. Fig. 19. ( X 39.5) : a Zellendc. h kleinlninige Zelle, c Liimcui mit Erweiterung (er), d durch Kupferoxydammoniak gequolle Ansicht, v Verschie- bung, st Stärkekörnchen, tr innere Haut. Fig. 20. {x?m): Querschnitte. Fig. 21 und 22. Die Bastzelle von Broussonefia l'aslnoJ:!. Fig. 21. (x395): a-c Längsansicht des IMitteltheiles. (t, handförmige Bast- zelle. I) dickwandige, c schmale lîastzelle. d Zellende, h UmhuUungs- schicht. Fig. 22. ( X 39.')) : Querschnitte. // wie ohen. Fig. 23 und 24. Die Bastzelle von Tilia cordata, var. japonica. Fig. 2.3. (x39.">): Querschnitte. Fig. 24. ( X 39.j) : (/ Längsansicht des Mitteltheiles. h, c Zellende, r w(>llig contourierte Stelle, p Porenspalten. Fig. 2;5 und 26. Die Bastzelle von Hihi so/s si/riac/is. l'ig. 2,j. ( X 39.")) : a Längsansicht, h, c Zellende, en Verengerung dt's Lu- mens, s Verzweigung der Ende. Fig. 2(]. (x395): Querschnitte. Fig. 27 und 28. Die Bastzelle von Firminia platanifoUa. Fig. 27. ( X 39.')) : a Längsansicht. /), c Zellende, c Zelle mit unregelraässigen Ausbauchungen der Wand. |) Porenkänale, en Verengerung des Lumens, er Anschwellung des Lumens, ?: Verzweigung. Fig, 28. (x39.5): Querschnitte. Fig. 29 und 30. Die Bastzelle von Abutilon Aviccnme. l'ig. 29. ( X 39.")) : a, h Längsansicht des Mitteltheiles. c. Zidlende. cn Verengerung des Lumens. Fig. 30. (x39;)): Querschnitte. Fig. 31 und 32. Die Bastzelle von Uhnus montnna, var. laciniato. Fig. 31. (x 39.')) : Querschnitte, an, in äussere und innere Schichte der Wand. Fig. 32. ( X 39.")) : n, b Tiängsansicht des Mitteltheiles. h Zelle mit Verschie- ])ung (?•) auf der innere Wandsehicht. c Zellende, an, in wie oben. Fig. 33 und 34, Die liastzelle von Urena Jolxtta. Fig. 33. ( X 39.3) : a Längsansicht des Mitteltheiles, h Zellende, en Ver- engerung des Lumens. Fig. .34. (x39.')): Querschnitte. Fig. 3.") imd 30. Die Ixistzelle von Wisfario cj/inens-is. r'ig. 'Î,"). ( X 39;")] : Querschnitte, j) Poreuspalten. a/t, in, äussere und innere Wand seh ich te. Fig. '.]('). ( X :>!).")) : o, h Längsansicht des Mitteltheiles. c Zcllende. /.• Kern, V Verschiebung, on, in, p wie oben. Tafel. XXI. Fig. 37-39. Die Bastzelle von Celadrus articulalus. Fig. 37. ( X 395) : Längsansicht des Mitteltheiles. a taiigentiiile Ansicht, b radiale Ansicht, p Forenspalten, v Verschiebung. Fig. 38. (x395): Zellende. Fig. 39. (x395): Querschnitte, p wie oben. Fig. 40 und 41. Die Bastzelle von Boelimeria nivca. Fig. 40. ( X 395) : a Längsansicht des Mitteltheiles. b Zellende, v Verschie- bung. Fig. 41. (x395): Querschnitte. Fig. 42-44. Die Bastzelle von Bœhmeria splcata. Fig. 42. a-f, Längsansicht des Mitteltheiles. a-e ( x 39,5)./ ( x 135). g Zellende ( X 395). en Verengerung des Lumens, cap Einkapselung des Lumens, h Kern, a, ß Zwei Erweiterungen des Lumens. Fig. 43. (x395): Erweiterung (er) des Lumens, st fc^tärkekörnchen. caj) wie oben. Fig. 44. (x395): Querschnitte. Fig. 45 und 46. Die Bastzelle von Curchorns capsularls. Fig. 45. ( X 395) : a Längsansicht des Mitteltheiles. b Zellende, cn Verenge- rung des Lumens. Fig. 46. (x395): Querschnitte. Fig. 47-49. Die Bastzelle von Viti-s Ooignetiœ. Fig. 47. ( X 395) : Längsansicht des Mitteltheiles. a primäre Bastzelle, b sekundäre Bastzelle, qr Querwand, p Porenspalten. Fig. 48. (x395): Zellende, a primäre Bastzelle, b sekundäre Bastzelle. pj wie oben. Fig. 49. ( X 395) : Querschnitte, a, b, p wie oben. Fig. 50-52. Die Bastzelle von Edgeworlliia papyri/cra. Fig. 50. ( X 395) : Querschnitte, bp Bastparenchym. Fig. 51. ( X 395) : Längsansicht des Mitteltheiles. a Zelle mit den durch Chlor- zinkjod hervorgerufene Querstreifen der Wand, b unregelmässig con- tourierte Zelle, c Zelle mit Porenspalten (p). d Zelle mit Verzweigung Fig. 52. (x395): Zellende, a mit Verengerung des Lumens (cn). b, c mit Verzweigung ;., Fig. 53-.56. Die Baytzellc von Caimahla t;ativa. Fig. 53. ( X 395) : Qiier«chnitte. au, in äussere und innere Wandschichk'. Fig. 54. ( X 305) : Längsansicht des Mitteltheiles. v Verschiebung, au, in wie oben. Fig. 55. ( X 395) : Zellende, a nuriual. b verzweigte. Fig. 56. ( X 395) : Innere Haut nach Behandlung mit Kupferoxydammoniak. Fig. 57 und 58. Die Bastzelle von Urtica Tliunhergiana. Fig. 57. ( X 395) : Querschnitte. Verdickungsschichte radial gestreift. Fig. 58. ( X 395) : a Zellende, h Längsansicht des Mitteltheiles. h Kern. Fig. 59. ( X 395) : ßastcambiumzelle von Urtica Tlmnhcrgiana. hc Bast- cambiumzelle. s.-s. Stärkescheide. Fig. CO. ( X 900) : Alle Stadien der amitotischen Kerntheilung. nuc. Kern- körperchen. Fig. 61. ( X 900) : a Zelle mit ruhendem Kern (/j). h dieselhe mit ruhendem Kern (k) und mitotisch sich theilenden Kern {ml:). Figuren 60 und 61 be;iiehen sich auf Bastcambiumzellen von Urtica Tlitmbergiaua. Jour. Sc. Coll. Vol. XV. PI. XX. K.Saito. d,]. Lith. B. KoibihB. . Tokvc Jour. Sc. Coll. Vol. XV. PI. XXI. K. s ai to. del Lith. E. Koshiba.. Tokyo Untersuchungen über die Schrumpfkrankheit („Ishikubyö") des Maulbeerbaumes. II. BERICHTE) VON M. Miyoshi, RigahihahisM. Professor der Botanik, a. cl. Kaiserl. Univ. z. Tokio. 1. Mein fj'iilierer Befund, d:is,^ dio Entloovnnn- {\ov Assiniilnto Ikü (Ich erkrankten, jedoeli noeli \r»lliL;' grünen l>l:ittern nni- unvollkonnnen Btatllindel, wurde dnreli die von Ende März \m Ende October 1900 alle 5 Tage ausgeführte Jodprobe bestätigt. Diesell)e Tliatsaclie wurde ferner durch Verdunkeluno'sver- suche in der Weise nachgewiesen, dass kranke Blätter auf intakten Pflanzen, welche entweder an ihrem natürlichen Standorte mittelst grosser schwarzer Pappc}dinder bedeckt, oder in Töpfe gepflanzt 1) Eine anslTilirlicIie Mittlieilnng lielintlet sieli in ja]>anis(:licr Siiraclie in einem anil- liohon „Eerielite über dieJSclirnnipfkranklieitJdes Manli>oerlianines," Bd. V. 1001. p. 4fi5 n. s. w. Der I. Berielit über die Resultate vorliejiender^Stndien erscliien im Botanisciien Centralblatt, Bd. LXXXIII. No. 11. 1900, ausführlicher aber im oben erwähnten amtlichen Berichte, Bd. IV. 1900. p. ISS n. s. w. 4C)0 M. MTYOSTTT : FNTERSUCITUNGEN UEBER DTE in oinon Dunkelranm gebraclit worden Avaren, ihre Assimila tions- Htürke 4 (xler ö Tage (in einigen Fällen über eine Woclie) naeli (lei- Vei'dunkelung noch behielten, während die Kontrollo1)jeete (normale Pflanzen) nnter denselben Umständen sieh schon nach einem bis zwei Tagen vollständig stärkefrei erwiesen. Dass diese schwache Entleernngsßihigkeit bei den kranken Blättern nicht etwa durch Diastasemangel verursacht ist, wurde schon früher iiervorgehoben^) und mm als richtig erwiesen durch eine lîeilie von Yersuclien die Herr K. Shibata speciell für den Zweck ausgefülirt hat. Die kranken Blätter ftmd er stets (ohne eine einzige Ausnahme in seinen mit 4 Kulturrassen des ^Maulbeerbaumes angestellten Yersuchsserien) reicher an Diastase als normale Blätter, als er seinem in Zimmertemperatur, in einem Falle bei 40°-50°C, zubereiteten Blattauszug Stärkekleister zusetzte und mittels üblicher Cu.jO-Messungsmethode nach dem Kochen mit Fehlingscher Lösung und auch mittels der Farben- reaktionsmethode nach dem Zusatz von Jod nachgewiesen hat. Dieser Befund zeigt unzweideutig, dass die Diastase der kranken Blätter ausserhalb der letzteren ihre volle Wirkung äussert, und schliesst von vornherein den Gedanken aus, dass die ungenügende Entleerung der Assimilate bei denselben Blättern durch den liemmenden Einfluss einer gewissen Inhaltsnbstanz (z. B. Oxydase,) bedingt sei, denn diese hätte mit der ^Diastase zugleich in den Blattauszug übergehend, dort auch ihre Wirkung zeigen müssen. Soweit mm meine bislan«' o-owonnenen üntersuchunoserffeb- nisse es erlauben, sei hier hervorgehoben, dass der fragliche Grund anderswo liegen nmss als oben gesagt ; er liegt nämlich, wie schon früher angedeutet,-) in den anatomischen Merkmalen der kranken 1) u. 2) Verg. meine Älittlioil. im anitl. IViiclitc ü. d. Schrmnpfk. d. Manlbecrh. Bd. IV. 1900 p. 21G. SCHKUJirFKKAXKHEIÏ (iSHIKüBYü) DES MAULBEERBAUMES. 461 Blätter d. li. der imvollstäiidio-eii AusbiUluui;' der Stoff k'ituiideii BjdiiK'u, (Irr Siübrölireuglieder. Die geringe Luiiieubreite der iiiiiiiliclien Leitbalineii, welclie liier überhaupt in geringer Anzahl vorhanden sind, gestattet nur eine äusserst langsame Wegführung der Assimilate, (hier speciell des Zuckers), infolgedessen die weitere Auflösung der Assimilationsstärke gehindert wird, was aus den Yersuchsergebnissen von Haust een, Pu rie wit seh und Lintz genügend bekannt geworden ist. Wird indessen die AYirkung der Diastase verhindert, so kommt doch infolge des durch Stärke- anhäufung im Chlorophyllkörper ausgeübten Reizes innuer wieder Neubildung des Enzyms zu Stande und somit resultiert der ol)en erwähnte Ueberschuss. Uns liegt hier ein interessanter Fall vor, welcher zeigt, wie anatomische Abnormitäten eine Reihe tief- greifender, physiologischer Störungen zu Folge haben. Da die Entwickelung des Blattes, wie Vöcli ting 'sehe Versuche uns lehren, von seiner Assimilationsthätigkeit abhängig ist, könnte die oben besprochene Beeinträchtigung der C- Assimi- lation möglicherweise auf das weitere Wachsthum der Blätter hindernd einwirken. Experimentelle Beweise über diese und andere wichtige Punkte werden fernere Studien bringen. 2. Eine Reihe (7 Serien) von Blutungsversuchen wurde von September Ijis Novend^er 1900 bei im Freien wurzelnden normalen und kranken Stämmen ausgeführt, und ergaben sich folgende Resultate : a) Der maximale Druck eines 3 jährigen, gesunden ca G cm Umfang habenden Stannnes (Kulturvarietät ,,Roso"), welcher an einer Höhe von 5 cm über der Ei'de ireschnitten und mit Manometer versehen wurde, wurde am 19. Septemljer erreicht und entsprach einer Quecksilbersäule von 70 cm ; ]>) der tägliche Maximumdruck wurde fast in allen Fällen ungefähr um 12 Uhr mittags erreicht ; c) kranke Stämme zeigten im Vergleich zu 462 M. MIYOSHI : UNTERSUCHUNGEN UE13ER DIE gk'ielijiilirigen, gleicligrossen, gesunden Stämmen derselben Kul- turrassen, stets geringeren Maximumdruck, (z. B. in einem Falle bei normalem 74 em und lu'i krankem 27 em, in anderem l^^alle l)('i normalem 'j4. em u]u\ bei ki'iinkem 7 em) ; d) in nlleJi Fällen sajdv der Blutiuiiisdi'uek nur nlhnälilieli lierab und in einiireii Fällen trat bald negativer Druek zu Tjige. Diis Wurzelsystem der erkrankten Objeete sali noeli voll- kununen normal aus und war fast ebensogut entwickelt wie l)ei den Kontrollobjecten, nur waren dickere Würzelclien bei den ersteren weniger zalilreicli als bei den letzteren. Mikroskopische Untersuchungen zeigen aber einen eklatanten Unterschied, dadurch dass bei den Kontrollobjecten sowohl dicke als auch dünne AVih'- zclchen ihre Holzcylinder mächtig entwickelt hatten, dagegen l)ei den erkrankten der Holztheil verhältnissmässig dünner und die Kinde dicker war. Ferner hatten die erkrankten Objeete eine geringere Anzahl von Gelassen, deren Lumen wiederum kleiner war als bei den kon trollen. Diese unvollständige Ausbildung der Wasserbahnen muss somit als der Hauptgrund angesehen werden, welcher in dem kranken Stamme einen w^eit schwächeren Blutungsdruck verursachte. Die Ti'autspirationsgrösse der beblätterten Zweige nach den Mitte September und Anfang October ausgeführten AVryuchen erwies sich l)ei deai erkrankten J^xemplaren als viel geringer bei den Kontrollobjecten — ein Unterschied welcher ebenfalls auf der Ausbildungsweise der wasserleitenden Flemente beruht, abge- sehen von der Beschaffeidieit der Blattepidermis, Function der Spalt(")fthungen u. s. w. ?j. Die alle 3 Woeben bei einigen Kulturvarietäten des Maulbeerbaumes ein ganzes dalir biiuhu-eli ausgeführten Messungen der Dicke des Holztheils ergaben, dass bei gleichdicken Zweigen SCHRUMPFKRANKHEIÏ (iSHIKUBYo) DES MAULBEERBAÜ3IES. 463 durchsclinittlicli bedeiiteiid weiiioer Holzbilduns; in crkraukteii Objceteii stattfand als Ijei den gesunden, und aueli die Stärkc- nieuge in verseliiedenen Tlieilen eines Zweiges l)ei erkrankten stets irerinircr wai* als bei 2;esunden. Dieses seliwaelie Dieken- waelisthuni ist eine Folge des Blattabpflüekens (Vei'g. Jost's scliüne Untersuchungen ühvv „Beziehung zwiselien Blattentwieke- lung und Getassbildung u. s. w." Botan. Zeit. 189o), und die geringe Zweigstärke beruht auf unvollkonnnener Ausführung der C- Assimilation . 4. Dass erkrankte Stämme, die in ihren anfänglichen Stadien von übermässigem Zweigschneiden und Blattabpflücken für einen gewissen Zeitraum verschont blieben, oft zeitweise sich erholen, zuweilen völlig geheilt werden können, ist eine unter jMaulbeer- l)aumzüehtern bekannte Thatsache. Um eigene Erfahrung darüber zu gewinnen, liess ich eine Anzahl kranker Stännne in einem Kulturboden von Sommer 1809 bis Herbst 1900 unberührt stehen. AVie erwartet, waren folgende Resultate vorhanden. Name der Kultvirra^sen dos Maiill>eerljauuies. GesaiiiDit- zahl der uiiheiührt «^-elasseneii kranken Suiniine. AbgesL(jr)ieii. ]\iank geblieben. A'ol Island ig geheilt. " Zalil der unlie- rührt gelasse- nen gesnnden Stiinnne (Kcntn.ll- Ulijeete.) Jumonji. Nezumigaeshi. 9 8 1 2 6 1 2« 5* 8 (Alles sresuuil geblieben.) 4 (AVie ubeij.) * 3Iäcliti,L;e Ildlzliildiiiig und reicldiciir Slärkeablagerinig wie liei Konlrcill-l'llanzen. 1) Verg. i. IJeiielit. I.e. nnd aiieli meine Mittheil, im aüitl. lîeriehle ii. Sehrnm]irk. d. Mauüieerb. Ed. IV. l'JOU. p. 'loS u. s. w. 4G4 Somit stelK'ii die <)1)l'ii ziisaniineiigcfassten Ergebiiisso mit (k'iijcnigeii, welche im I. Berichte vorliegender Untersuclmiigeii mitgetheilt waren, völlig im Einklang und bestätigen meine über die Ursache der Öchrumpfkrankheit früher geäusserte Annahme. Weitere einschlägige Versuche sind im Gang. Untersuchungen über die niederen Organismen welche sich bei der Zubereitung des alkoholis- chen Getränkes „Awamori" betheiUgen. YOX T. Inui, EigahisJii. Mit Tafel XXIT. Ein stark alkoholhaltiges, dem Whisky ähnliches Getränk ,,Awamori"^) wird schon seit vielen Jahren auf den Lnchu Inseln') hergestellt, wo dasselbe einen wichtigen Handelsartikel bildet. Obgleich eine authentische Angabe über die Zeit fehlt, in welcher dieses Getränk zum ersten Male dort gebraut wurde, so ist doch wahr- scheinlich, dass man die Braumethode vor ca. f500 Jahi'en von Chinesen lernte und bei Izumisaki, einer A1)theilung der Stadt Nawa, die Herstellung versucht wurde. Gegenwärtig bildet die Stadt Shuri den Hauptsitz der Awamorifabrikation. 1) „ Awainori " bedeutet Scliaumweiii. 2) Luclin ist eine Itriel-Grnppe welche zwisclien Formnsa und den Kiiislin-Inseln liegt. 406 T. TNÜT : UNTEES. Ü. D. NIEDEREN ORGANISMEN WELCHE SICH Braumethode. Die BrnumetlioJo des „Awamori " bestellt ans drei Opern- tionen : 1) Die Bei'eitung des ,, Kqji." 2) Die Darstellung des ,, Moromi." 3) Die Destillation. 1) Die Bereitung des ,, Koji." — El)ens() wie 1)eim ,, Sake" so ist auch lu'iin ,, Awamori" Koji eine durch Vegetation eines Pilzes auf gekochtem Beis ge1)il(lete Masse. Der im Awamori- Koji hefindliche Fadenpilz ist aber von demjenigen des Sake-Koji verschieden und besonders durch seine schwarzen Sporen aus- gezeichnet. Die Technik der Kojizubereitung ist folgende : man wascht zunächst 82 Litei- geschälten Beises, lässt ihn 12-15 Stunden in AVasser liegen, wäscht noch mehrere Male und dämpft ihm dann 0-4 Stunden, worauf er in eine Kojihütte gebracht, und auf Strohmatten ausgel)reitet wird. Sodann ninunt mau eine kleine Menge dieser Beismasse, mischt sie, wenn sie sich l»is (10'^ TO^C/'^) abgekühlt hat, mit 2 Deciliter Tanekoji") und (Heses wiedei" mit dem ü])i'igen Beis. Wenn der gekochte Beis sich abküldt und eine Temperatur von ca 30°C erreicht, l)edeckt man ilni mit Sti'ohmatten, um ihn von weiterer Abkühlung zu schützen und ihn möglichst feucht zu halten. Nach 12 Stunden ist die MvcelentwickelauLr zu seilen und 1) Olif^liicli (liest- 'PcmiHTiitiir (Kr ErlKiltuni; dfs Lehens der Sporen unt;iins(ig zn sein scheint, überleben doeli in WirUlielikeit die meisten Sporen da der (•ekoehine Reis sehr rasch sieh :vl)külilt. Experimentell hal)e ich sogar gefunden, dass die Sporen dieses Pilzes gegen die Wiirnie sehr resistent sind. Parüi)er wird bald die Rede sein. 2) „ Tane-koji " ist nicht anders als auf Hirse kultiviertes Sjjoren tragendes Myeel. B. D. ZUBERETTTTNG D. ,,AWA:\rORl" BETITEILTCEX. 407 (lie Toniporatur steigt l)is 29°C 0)ei 27°C Lufttemperiitur). Naeli A'ei'laiit' des orston Tages steigt die Temperatur des Koji bis 32^C lind einige schwärzlichen Sporangien werden sichtltar. Am divitten Tage ist der H(ihepunkt der Pilzentwickelung erreicht und die Temperatur des Koji ist ca o4°C. Dann nimmt man die Matten fort und lässt die Temperatur nicht weiter steigen, was für die Beschaftenheit des Koji schädlich wäre. Am 4ten Tage werden die Hyphen gänzlich reif, die Oberfläche des Koji bedeckt sich mit schwarzen Sporen und die einzelnen E-eisk(')rner werden nun von den Hyphen mit einander zu zahlreichen Kläm])chen verbunden. Die Entwickelung der Hyphen aitf Awamori-Koji ist nicht so üppig wie beim Sake-koji, da die dort benützten Räumlich- keiten viel Feuchtigkeit entweichen lassen. Sinkt die Lufttem- peratur bis zu 12°-13°C, so entwickeln sich die Hyphen nur langsam und das Koji wird erst nach l-j-lß Tagen fertig. 2) Die Darstellung des „ jMoromi.'*— Man gielit 82 Liter Koji in einem Bottich mit 7o Liter Wasser und 4 Deciliter ,,Tane- Moromi " (schon in Gärung begritfene Kojimasse) und l)edeckt diese Mischung mit einem grossen ])eckel. Nach 3 Stiind(Mi zeigt sich l)ereits die Gärung durch C^0-2-(ias Entwickelung an. Jetzt steigt die Temperatur im Bottich allmählich und übertrifft die der Luft. Am dritten Tage erreicht die Gärung den höchsten Punkt, wobei die Temperatur der Gärmasse manchmal ?A°C erreicht. Xuii sinkt die Temj^eratur wieder, und am 8 ten Tage hat sie gleiche Höhe wie die der umgebenden Luft. Bei allzu niedriger Tempei'atur verläuft der Gärungsvorgang nur langsam und die Beife des Moromi braucht eine lange Zeit. Folorende zwei Tabellen entnehme ich aus meinen Beobacht- 408 T. INUI : UNTEPtR. Ü. D. NIEDEREN OEGANISMEX WELCHE SICH iing'on, uni dio tngliehon Scliwanknngon dor Tenijx'rntur ini Bottiche zu vorniiscliauliclieii. Tal .elle L (Bei 27°C Lnftten.pcratur.) 1 stell Tag 28° C 2 ten Tag •^9° j) 3 >j >; 32° j) 4 j? }} 30° 5! 5 V }> 29° 5> 7 1} >! 27° >> G »5 }) 27° n S ;> }) 24° 5> 9 }> }> 24° l> 10 9Î ?) 23° )) IL ten Tag 23°.5 C, 12 J7 j> 24°.0 „ 13 )> )) 23° 5 14 !7 V 23°.8 „ l.") V >! 23° 5 IG Ji }; 17 }} T> •>3° S LS )} )> 23° 8 19 » ■>■) 23°.5„ 20 )> )> 23°.5 „ Tal .elle IL (Bei 21°C Lnfttemperatnr.) 1 sten Tag 22° C. 11 ten Tag 17°.8 C, 2 ten Tag 24° „ 12 j> }) 17°.5 „ }} 24°.5 „ 13 » v 17°.5 „ 4 >} :24°.0„ 14 >j }} 17°.8„ 5 >} 23°.0 „ V) J! }} 17°.8 „ G 7} 23°.0 „ IG ') 77 17°.5 „ 7 57 21°.5 „ 17 ;» 77 17°.8 „ 8 5> 18°..5 „ 18 >' 7» 17°.5 „ 9 J) is°.o„ 19 5' 77 L7°.0 „ 10 )J 17°.5 „ 20 ,, 77 17°.0„ Erst nach 17 odoi' LS Tagen im Sonimor und nach oO Tagen im Winter unterwirft man den Bottich-Inhah der Destillation. :>) Die J)estillatiou. — Die Destillation wird sehr einfach aus- geführt und l)raueht hier nicht geschildert zu werden. Zu dem Destillate setzt man luni geröstete Hirse, welche man lange Zeit darin liegen lässt, wodurch bewirkt wird, dass der „Awamori" B. D. ZUBEKEITUNG 1). „AWAMOlll" JiEïllElLIGEN. 4()U hciin Ausgiesscii eine lebhafte Scliauiiil)ikluiig zeigt, was eine Eigeiitliiiniliclikeit des Ciretränkes ist. 2. Die Fadenpiize im Awamori-Koji. 1) Aspergillus luchuensU iiov. sp. Dies ist der wielitige Fadenpilz, dureli dessen seliwarze 8poren dem Koji sein eliaracteristisclies Aussehen verliehen wird. Er ist in reinem Koji stets vorhantlen und bedingt die Verzuckerung bei der Herstellung des ,,Muromi." Alle anderen Fadenpilze, die sieh oft in Koji vorfinden, sind nur zullillige Vorkonnnnisse. Morphologie : — Auf festen sowie in flüssigen Substraten bildet unser Pilz einen dichten verfilzten Rasen, auf welchem schon nach einigen Tagen zahlreiche, vertical emporsteigende, kurze, weise Konidienträger erscheinen. Bald darauf bilden sie weisse Köpf- chen, die sich dann allmählich vergrössern und zugleich bräunlich färben. Nach etwa 3 Tagen werden sie schwarzbraun und die Konidienträger verlängern sich bis 2-2.5 cm, deren dickwandige, feste, farblose Stiele noch ganz glatte, kugelige, schwarzbraune Köpfchen tragen. Später nimmt die Oberfläche des Köpfchens ein durch massenhaftes An^Yachsen der Sporenmenge bedingtes un- regelmässiges Aussehen an. Der junge Konidienträger wird sehr leicht mit blossem Auge in seiner Form erkannt, wie es auch bei Aspcrylllus WeiUu,'') und Aspergillus (jlaucus der Fall ist. Auf gckochotem Keis gedeihen die Hvphen üt)t>ig und erzeugen reich- lich Mycel. Man findet aljer darin nicht die langen, verzweigten, aufsteigenden Hyphen, wie man sie beim Asp. Wentii sieht. Der Mycelfaden erreicht manchmal 8 p- Dicke, die Wand ist dünn, 1) Well Hier, Centrallil. 1". Dakt. IM. II. p. 140. 470 ï. INUl : UNTEllS. Ü. D. NIEDEREN ORGANISMEN WELCHE SICH Yerz\vei2;t sich an verscliiedenen Stelk'ii und zeiirt mclirmals Quer- tlieilunii;. Diu Hy[)lu'n, welche über zwei Monate \nui^ auf testen kîu))- sü'aten i;elialten wurden, zeii;en lokale Anseliwellungen, indem das IMasnia allmiihlieh sich an gewissen Stellen ansammelt und dabei ein granuliertes Aussehen gewiiuit. liier entsteht die Scheidewand mid ein kugehges Gemma. Die Blase ist gewühnlicli kugelig, selten kolbenförmig und steht auf dem Stiele senkrecht wie bei A.^p. Wentii. Nach dem Ablalle der Sterigmen zeigt die Blasenober- Häche vielen polygonale Vertiefungen. Die Stielmendjran ist glatt, 1-3 « dick, farblos, nur bei alter Kultur wird sie braun. .Vuf de]- Blase stehen zahlreiche, radial ausstrahlende Sterigmen dicht an einander, und erzeugen Konidien- ketten an der Spitze. Die Sterigmen sind länger als bei den an- deren Arten, und haljen eine Länge von Y2-V3 Blasendurch- messer. In Gestalt und Grösse sind alle Konidien fast gleich- massig, sie sind 4-5 /i gross, fein warzig und kugelig. Elliptische Konidien wie 1 )ei Asp. Wentii und bei Asp. Oryzœ konnnen hier niemals vor. Perithecien konnte ich weder auf flüssigem noch auf festem Boden beobachten. Physiologie : — 30°C-'3ö°C ist die Optimumtemperatur für die l^ntwickelung der Hyphen. Bei lb°C vegetiert der Pilz sehr langsam ; und bei 12°C geht keine Sj^rossung mehr vor sich. Der beste flüssige Nährboden scheint das Koji-lvxtract zu sein; gut ist auch die mit l-^o Trauben- oder Bohrzucker zugesetzte Kaulin'schc Nährlösung, in welcher die Konidien schon nach 24 Stunden l)ei 25"-28°C sprossen können. In Biei'würze gedeihen die Hyphen ebenfalls üppig und l)ilden Konidien erst nach 20 Tagen. Unter den festen Nälirl)öden ist der gekochte Beis am günstigsten. Brod, AVih'zegclatine, Pleischpeptongelatine mit Zucker, Zui;ker E. D. ZUBEREITUNG U. „AWAMOßl" BETHEILIGEN. 471 Gelatiiu' mit Nälirsalzeii sind ebenfalls gnt. Auf festen ^»'älirbödeu gellt im /Vllgemeinen die Entwiekeluni;- der Hyplien sehr rasch von statten und der Konidienträger wii-d auch friÜier gelnldet als in flüssigen. In Stärkekleister, weleher ausser 2'?ö KartoHelstärke noeh die nöthiocn »Salze enthält, findet Verzuekeruni;- nur lauirsam statt. Die Gelatineverflüssigung ist bedeutend ; bei schiefer Kultur ist die erste Verflüssigung scliun nach 4 Tagen wahrnehmlja]', und nach 30 Tagen vollständig. In Bezug auf die Widerstandfähigkeit der Sporen gegen höhere Temperatur habe ich einige Versuche angestellt und gelangte zu folgendem Kesultate ; bei 1-stündichem Erwärmen auf 60°C behielten die Sporen noch ihre Keimfähigkeit, während nach eljenso langem Verbleil)en bei 7ifC die Keimung nicht mehr statt- fand. Ein kleiner Theil der Sporen zeigte aber Entwickelung, wenn dieselben bei 70°C nur eine halbe Stunde lang erhitzt w^orden waren. Somit besitzt unser Pilz einen hohen Grad von Resistenz gegen Wärme. Was die Sprossungsschnelligkeit der Konidien betrift't, so konnten wir keinen Unterschied zwischen der jüngeren Kultur und der über 2 Monate älteren Kultur beobachten. In Bierwürze entstehen die Hefe-Genniien nicht und tritt keine Gärung ein. Vergleich mit ähnlichen Arten: — Aspergillus Tleyi^// Welimer, welcher mit der vorstehenden Art viele Aehnlichkeit besitzt, unterscheidet sich dadurch, dass er bei Eeagenzglaskultur aufwärts emporwächst und verzweigte Hyphen bildet, während A. luchuensis, wie A. Oinjzce, auf der Kulturfläche viel kürzere Luftmycelien entwickelt. Ausserdem Ideiben die Hyphen des A, luchuensis das ganze l'^ntwickelungsstadium hindurch farldos, AV^as die Earljc der Konidien betrüft, so verändert sie sich bei 472 ï. INUI : UNTERS ï'. D. niederen ORGANISMEN WELCHE SICH A. Wentil von grüngelb zu Lrüiuilichgelb, Lei Ä. liichucnsis dagegen von wei^.s zu dunkelbraun und 8cliwarzbraun. A. Menlil produciert oft elliptische Konidien, während A. luchuensU nur rund liehe. Hinsichtlich der ()])tinuunteniperatur für die Hyphen- Entwiekelung existiert el)enfalls ein gr(jsser Unterschied zwischen den beiden .Vrten. Nach Welimer\) wäelist A. Wejitii l)ei lo°-lb°C am besten, trotzdem er Bewohner eines tropischen Klimas ist. Dahingegen ist oÜ°-3ü°C als die günstigste Temperatur für ^[. /^^t'Ä«<3;^5/s erwiesen, und bei 12°-13°C findet das Wachstum überhaupt nicht statt. Diagnose. Steriles Mycel weiss, mit mehreren Septen versehen, stark \erzweigt und dicht verflochten. Konidienträ.ger kurz, dicht stehend ; die Stielmendjran glatt, durchsichtig. Köpfchen zuerst weiss, dann hellljraun, endlich schwarzbraun. Blase glatt, kugelig, oft oval und zeigt nach dem Aljfalle des Sterigmens dreieckige oder pülygüJiale Vertiefungen auf der Oberfläche. Öterigmen lang, radial ausstrahlend. Keife Konidien kugelig, fein warzig, 4-4. ~) ,"■ in Durchmesser. Perithecien fehlend. Die Optinuun-Temperatur fin- Hyphen-Entwdckelung ist 30''-3ü''C'. Gute Nährboden der Kei- henfolge nach sind Keis, Brod, Ca'laüne. Gelatine Verflüssigung bedeutend. ( i rü.ssciiVL'rlüU üiiss. Jlv"]iliL'iuliirciiiiU'.s,ser 2-8 //. Konidien träger 1-2 nun ludi. ISticldicke JO-I.j/a Kopfchendiuchnies.ser 40-80 //. Blafendurclimesper 20-30 //. Htcrignuni i> ff. x o //.. KoriidiendnrulnncHser. 4-4.,") /l J) Weh nie r, 1. c. p. MI. p.. I). ZUBEREITUNG D. „AWAMORl ' r.ETlTEILTCEN-. 4/-) 2) Aspergillus^ perniciosus nov. sp. Dieser Pilz ist aueli ein liäiifiges Vorkommiiiss in Awanioi'i-Kqji und zeielmet sieli dureli seine branngelb gefärbten Sporen aus. Obgleich er in gutem Koji nur wenig oder nicht vorhanden, gelangt er oft zu beträchtlicher Entwickelung und verdrängt j\. luchuensis. Sein Verznckerungsverniögen ist schwächer. Die Hyphen sind gelbgrünlich gefärl)t ; die Köjifchen sind anfangs weiss, dann gelb, und schlieslich graubraun. Der Koni- dienträo'er beträo;t 2.5 mm in lunsfe. Auf o-ekochten Eeiskörnern entstehen niemals senkrechte, verzweigte Hyphen, ein Merkmal, welches den Pilz einerseits mit Ä. luchuensis in nähere Beziehung bringt und andererseits von A. Weiitii unterscheidet. Das Köpfchen dieses Pilzes ist im Verhältniss zum Stiel bedeutend grösser, als es bei A. luchuensis der Fall ist, und in dieser Beziehung ist unser Pilz mit A. nieder verwandt. Sterigmen sind etwas kürzer als Ijci A. luchuensis und A. Wentii, und überschreiten niemals /:; des Blasendurchmessers. Die Konidien sind warzig, kugelig, ihr Durchmesser misst 4-5 /^ Perithecienbildung fehlend. 3) Monilia sp. Einer Monilia-Art l)egeguet man clx'ufalls in Koji und un- gleich der 3Ionilia variabiris Lintner ändert sie ihre Form nicht bedeutend. Die Hyphen, die sich in Bierwürze von ein- zelnen Zellen entwickeln, bilden radiale Kolonien, deren Centrum etwas undurchsichtig wird und nach 5-0 Tagen zeigt die (Jber- Üäche der Kolonien einen weissen staubartigen Auldick. Die Kolonien auf Fleischpepton-Gelatine zeigen jedoch nicht radiale Anordung der Hyphen wie im obigen Falle ; ihr Band bildet ferner ganz unregelmässige V^orsprünge. Die Fäden sind mehrmals septiert, und verzweigt. Beim Abschluss der Luft werden auf der Oberfläche eine Menge kleiner ovaler Sprosse gebildet. 474 T. TNUI : UNTERS. V. D. NIEDEREN ORGANISMEN WELCHE SICH welche sich abtrennen, nnd wie Hefepilze fortpflanzen. In Bier- würze findet eine schwache Gürnng statt nnd nach lô Tagen sind 3?^ Alkohol in 1 Liter Würze entstanden. Die Hv])hen verflüssiiren die Gelatine. 3. Die Sprosspilze im Awamori-Koji. 1) SaccJud'Oiiiyce^ Awamori nov. sp. Die vorstehende Art ist eine gärtüchtige Hefe von Awamori und lässt sich nur in Moromi, nicht aber in Koji, anffinden. Auch in der Lnft der Awamoi'ibranerei ist sie nicht vorhanden, wo ich sie mehrmals vergeldich gesncht habe. Ihr Ursprung ist somit nnbekannt, sie findet sich immer im ,,]\l<>romi," welches als Tane- moromi (Gärmntter) seit jeher von Bottich zu Bottich übertragen worden ist. Auf Biei-würze-riattenknltur haben die Kolonien einen kreis- runden, glatten ITmriss und eine centrale Vertiefung. Nach 10 Tagen wird abei- der Band nnregelmässig zackig, indem vom Centrum nach der Peri])herie eine Anzahl von radialen Falten aus- strahlt. Die Zellen sind anfangs elliptisch und nehmen später eine rundliche Form au. Bei der Bierwürzekultui' sind sie meist ellip- tisch, dagegen auf Zucker-Agar rundlich. Weder bei 30° C nacli Ablauf 24 Stunden noch bei 13°-ir)°C nach 3 Tagen tritt 8]iorenbildung ein. Gegen Wärme äussert sie einen gross(Mi AViderst;md, ein drei- stiiudiges l'i-wärmen auf -"iO^C vernichtet die lOutwickelungs- trdiigkeit noch ni(dit, erst l»ei (iO^C erfolgt dieses. (Jegen Alkohol verhält sie sich folgendermasseu ; die Fort- pflanzung der Zellen wird nicht beeiuflusst in einer 8?^ Alkohol- B. D. ZUBKREITUNC, D. „AWAMORl" EETHETLTGEN. 475 hnltigon Fliissigkoit, ho\ 13^^ Alkoholgehalt wird dio Eiitwickcluiig deutlich gehindert und noch deutlicher l)ei lo'?^, und endlich l)oi 20^0 hört die Eutwickeluug völlig auf. In dieser Beziehung scheint diese Hefe also weniger widerstandsfähig als Sakehefe. Tu Bierwürze kann sie ca G^/'o Alkohol in Volumen jiroduciei'cn. 0) Eine Form des Sacchoromyces anomrdus. Diese Hefe ist auch in Koji reichlich vorhanden, und ver- Icilit dem Awamori sein eigenthündichcs Aroma. Kolonien auf der Bierwürzjdattenkultur erscheinen anfangs als kleine Piniktc*hen, welche sich nach der Peripherie hin in ziemlich langsam, nach o])en aber sehr rasch, sich vergrössern, so dass sie sich endlich zu Stäbchen verändern, welche sich dann dui'ch eiirene Schwere allmählich nach unten bieo-en. Die Zelle ist kurzelliptisch, 3-5 ,« lang. Das durchsichtige Plasma enthält einige stark lichtbrechende Granula. Die Haut- bildung ist eine träge ; bei 30°C tritt sie erst nach 24 Stunden ein, und bei 14°-1;")°C nach 15 Tagen. Die Gärung in Bierwürze ist schwach mit reichlicher Bild- ung von Obstäther und dentlicher Säurebildung. Sporen werden bei 30°C nach 10 Stunden gebildet, sind Init- förmig, gewöhnlich drei. Resume. 1) Awamorikoji wird aus Eeis oder Hirse zubereitet. Die Entwickelung oi Eeagenzglaskultur zcigou die Lnftinvcclieii einen bedeuteiideu T'literseliied. Aucli die Optinuimtemperatiir für die Entwickehmg ist verscliieden. 3) In ^Vwaiiioi'i-Koji l)efiiidet sieh noch eine Art Fadenpilz, A. pmiiciosus, nov. sp., der A. lucliuensis sehr älmlieli ist. Die Sporen dieses Pilzes lia1)en anfangs eine grüne Farbe, wie bei A. luchueuM^. Der voi'liegende Pilz kann vniter Umständen dic! Kntwiekelung des A. hichuensi^ hindern. 4) Die wichtige Plefe für die Awaniorigärnng ist Sacrharo- )/iyees Awamori. Derselbe entwickelt sich leldiaft im Gärbottich nnd kann Q)^/o Alkohol l)ilden. Das eigentliche Aroma des Awamori l)ernht auf dem A"or- handeusein des Saccharoviyceft anomalus. Diese Untersuchnngen wnrden auf Veranlassung meines verehrten Lehrers Herrn Prof. Dr. Miyoshi während eines mehr als zwei-monatlichen Aufenthaltes (Januar ])is März, 190]) in Okinawa (Lnchu) an Ort und Stelle ausgeführt. Es ist mir eine angenehme Pflicht, am Schlüsse dieser Arl)eit ihm ineinen verlnndlichsten Dank auszusprechen. Herren Seminar-Direktor K. Ando nnd Seminar-Lehrer S. Kuroiwa in Okinawa bin icn für ihr während meiner Arbeit stets erwiesenes Wohlwollen nnd Interesse el)enfalls zu bestem Dank ver]'»flichtet. Ln Juli, 190L TAFEL XXII. Tafeler klär ling. Fi.s;. 1. Ein reifer Konidientriiger von Aspergi7li(s iHchuensis. Vergrössernng ca X 50. Fig. 2. Derselbe im Glycerin. Optischer Dnrchschinitt, Verg. ca x r)0. Fig. o. Derselbe in einem jüngeren Stadium. Yerg. ca x 200. Fig. 4. Eine kolbenförmige Blase. Verg. ca x 200. Fig. ,5. Dieselbe in einem jimgeren Stadium, nivcerinpräparat aus Reislvul- tur. Verg. ca x 200. Fig. 6. Konidien, drei von ihnen mit Keimscbläuclien. Veg. ca x GOO. Fig. 7. Gemmenbildung in angeschwollenen Hyphen. Würzgelatineknltur. A^erg. ca x 600. Fig. 8. lleagenzglaskultur von Aspergillus hrch/iensis auf gekochotem Eeis. Natürliche Grösse. Fig. 9-12. Konidienträger des Aspergillus perniciosus in verschiedenen Entwickelungsstadiura. Verg. ca x 200. Fig. 13-20. Verschiedene Formen von 3fonilin sp. ]3. Auf Wiïrzegelatine. 14. 15. 16. 17. In Würzelösung. 18. 19. 20. vVuf Wiirzegelatine mit Luftahschluss. Verg. cax600. Fig. 21-23. Kolonien von Sacdiaromijccs Aioainori auf Würzegelatine, in nacheinander folgenden Stadium. 21. 3-Tage alt. 22. 8- Tage alt. 23. 15-Tage alt. Verg. cax50. Fig. 24. Verschiedene ï'ormen von Sacchnromi/ces Aicamori : Verg. ca x 600. Fig. 25, Soccharomi/crs anomalns. Jour. Sei. Coll. Vol. XV. PI. XXII. LUh F. Koshiba Tokyo MBl, WHOI I.IHKARY yH iiKfl 1 ^7 ^ ■•■v.i; ilililiiill :r,v;..liii|l| &^ > ■ 't t'i'ih'rt-: ''■'(■A ;v;, '-'••■■; J* il;!'!!!!;;! i '