"m ^ ^ ^ ^ :^ m n as * ^ 'V. «i- THE JOURNAL OF THE COLLEGE OF SCIENCE, IMPERIAL UNIVERSITY OF TOKYO. Vol. XXXVIII. M -^ "^ m ± m w n PUBLISHED BY THE UNIVERSITY. TOKYO, JAPAN. 1916-1917. TAISHO S-6. Publishing Committee. Prof. J. Sakurai, LL. D., Rigakuliàkushi, Director of the College {ex officio). Prof. I. Ijima? Pa. D., RigalcuhakusU. Prof. F. Omori, Rigdkuliaknslii. Prof. S. Watasé» P^^- D., RigahuMhisM. IX'i^ I CONTENTS. Art. 1.— Über die Ausflussmenge des Elutungssaftes einiger Bäume. Mit 4 Textfiguren. By M. Miyoihi.— Pnbl. February 25th, 1916. Art. 2. — A monograph of Japanese Ophiuroidea, arranged according to a new classification. AV^itb 7 plates. By H. Matsumoto. — Publ. March 31st, 1917. Art. 3. — The great eruption of Sakura-jima in 1914. AYith 24 plates and 49 textfigures. By B. Koto. — December 25th, 1916. Art. 4. — Contributionen novae ad floram bryophyton japonicam. Pars .secunda. With 42 textfigures. By S. Okamura. — Publ. July 25th, 1916. Art 5.— On some Japanese calcareous sponges belonging to the family Heteroplidae. With 2 plates. By I. Hôzawa. -Publ. September 30th, 1916. Art. 6. — Injection Experiments on Plants. With 2 plates. By Y. Yexdo.— Publ. May 15th, 1917. 1 \- y s « PRINTED BY THE SANSHU-SHA, TOKYO. JOURNAL OF THE COLLEGE OP SCIENCE, TOKYO IMPERIAL UNIVERSITY. VOL. XXXVIII., ARTICLE 1. Über die Ausflußmenge des Blutungssaftes einiger Bäume. Manabu MlYOSHI, 2Ugalvs]ti, IiigaLiÜLukush i. Professor der Botanik an der Kaiserlichen Universität zu Tokyo. Mit 4 Textfiguren. Unter den in Japan einheimischen Banmarten zeigen nach meinen Untersuchungen Conms controvei'sa He:\is.^^ und Carp'mus yedoensis Maxim, einen hohen Blutungsdruck, welcher bis zwei Atmosphären erreicht.'^ Gleichzeitig mit den Messungen des Blutungsdruckes wurde auch die täghche Menge des Blutungssaftes, welcher aus einem an der Stammbasis gemachten Bohrloch ausfloß, beobachtet. Die Hauptergebnisse, die ich im Folgenden mitteile, stimmen einerseits mit den Resultaten früherer Forscher überein, zeigen anderseits aber einige Eigentümhchkeiten unserer Versuchspflanzen. Die beiden Baumarten, mit denen ich experimentierte, stehen 1) Nach der in neuerer Zeit zuerst von Hemsley (Botanical Mazagine T. 8261, 1909) vorgeschlagenen Benenmmg sollte unser Baum, welcher undeutlich alternierende Blätter besitzt, statt macrophylla, coniroversa heißen. L^ra Verwechslung zu vermeiden ist unsere Püanze mit dem neuen Namen bezeichnet. 2) tiber den Blutungsdrack bei Cornus macrophylla vergl. Miyoshi, Botan. Centralb. Bd. LXXXni No. 11. 1900; Ber. Deutsch. Botan. Ges. Bd. XXYH. p. 457. 1900; Ann. d. Jard. Botan. de Buit. 2. ser. suppl. UI. p. 97. 1909. 2 AET. 1. — MAN ABU MIYOSHI : im Botanisclien Garten der Universität zu Tokyo. Der Umfang des Stammes ca 50 cm über dem Boden mißt bei Cornus 1,66 m und bei Carpuius 2,35 m. Das Bohrloch, welches ca 2,3 cm breit und ca. 5 cm tief war, befand sich bei beiden Bäumen ca 50 cm über dem Boden. Das Loch wurde üblicherweise mittelst eines durchlöcherten Gummistöpsels fest verstopft, und durch den letzteren ging ein Glasrohr ins Loch hinein, um den ausfließenden Saft weiter durch ein Gummirohr ins große Meßgefäß abzuleiten. Um den Verscliluß zu sichern, wurde die Ansatzstello des Gummi- stöpsels mit einer Schicht Paraffin-Kolophonium überzogen. Das Zeichen des ersten Saftausflusses kann man in Tokyo bei Carpinus schon am Anfa-ng oder in der Mitte Januar und bei Cornus am Ende desselben Monats beobachten. Die Zeit ist aber von dem Standorte des Baumes abhängig ; an besonnten Stellen wird sie früher eintreten als au schattigen. Die tägliche Zunahme der Ausflußmenge im Januar und Februar ist kaum zu erkennen, vielmehr zeigt sie durch äußere Einflüsse bedeutende Schwankungen. Erst mit dem Eintritt der Frühlings Witterung (höhere Temperatur und größere Nieder scliläge) im März nimmt der Ausfluß aufflillend zu. Gegen Ende März oder Anfang April erreicht die Ausflußmenge ihreii maximalen Wert, indem sie bei Cornus auf 17 1 und bei Carpinus auf 40 1 pro 24 Stunden steigt.^^ Kurz nach dem Maximum nimmt die Menge ab, bis der Ausfluß gewöhnlich Mitte April vollständig auf- hört. Durch die Einflüsse der Witterung und andere Umstände, vornehmhch Verstopfung (siehe unten), zeigt der Kurvenverlauf der Saftmenge manche Unregelmäßigkeiten. Bevor ich in die Erörterung störender Ursachen eingehe, ist 1) Über die Ausfluß mengen von Cornus unJ Carpinus habe ich bereits in meinen in japanischer Sprache geschriebenen Handbüchern, „Die Pflanzenwelt Japans," 1910. p. 44, und „Vorlesungen über Botanik," 4. Aufl. Bd. I. 1911. p. 5i8, einiges mitgeteilt. ÜBER DIE AUSFLUßMENGE DES BLUTUNGSSAFTES ELSIGER BÄUME. 3 es angebracht als Beispiele die Ausflußmengen der Jahre 1911 und 1914 in folirenden Tabellen anzuheben. TAB. I.^^ Austiaß menge in Litern Temperatur'^) Relative Feuch- tigkeit ^0 Witter- ung Zeif^) Cornus con- iroversa I. Banm Cornus con- troversa n. Baum Carp'mus yedoensis Maxima G Minima C Bemerkung 23. L 1911. etwas etwas 0.550 yo -1.8° 50.7 Mar 21 0 Ü 0,175 3.9' -1.1° 52.6 „ 25 0 0 0.866 8.7' -3.5° 07.9 i. nachm. 26 0,039 0,026 2,700 iLr -2.2° 62.2 trül> 10. Torm. 27 4. nachm. „ 0,002 0,009 0,440 0,830 8.4' 1.4° 59.7 " Summe 1,270 28 0 0 0 3.7° 0.4° 91.8 Regen 29 0,063 0,021 1,840 12.4° 3.6° 82.5 trüb 30 0.004 0 0,260 11.2° 2.4° 81.3 " 31 0 0 0 7.6° 5.0° 92.3 Regen 1. n. 0,005 0 0,031 10.5° 4.0° 9Î.3 trüb 4. nachm. 2 0 075 0 0,776 14.3° 3.4° 73.4 Mar 3 0 0 0,088 10.0° -0.4° 53.7 „ 4. nachm. i 0 0 1,036 8.7° -1.8° 53.4 Mar 5 0,001 0 1,125 10 3° -1.8° 62.0 „ 6 0 0 1,050 10.1° -2.1° 56.6 Um 1.30 naclira bei Cornus lu. H 4. nachm. 7 0 0 0,588 8.8° -3.8° 59.3 ,, Lö:Uer tiefer gebohrt 4. nachm. 8 0 0 1,350 9.8° -2.2° 65.6 trüb 9 0 0 0,172 6.5° -0.1° 45.0 klar 10 0 0 0,119 8.3° -3.4° 5G.1 " Der Saft mehr oder M-eniger getrüb'; 1) Die Messungen der Ausflußmenge ^^-uraen von mnnem früheren .Assistenten Herrn Db. K. KoRiBA übernommen. 2) In dieser und den folgenden Tabellen ist dia Saftmenge von Mittag zu Mittag angegeben, sonst ist die Zeit der Messung stets erwähnt. 3) Die Ziffern der Temperatur und Feuchtig'^eit sind aus dçm amtlichen Wetterberichte des Zentralen Meteorologischen Observatoriums in Tokyo entnommen. ART. 1. — MANABU MIYOSHI : Ausflußmenge in Litern Temperatur Eelative Feuch- tigkeit 0/ /U Witter- ung Zeit Cornus con- iroversa I. Baum Cornus con- troversa U.. Baum öarpinus jedoensis Maxima C Minima C Bemerkung 11 0 0 0,158 16.1° -1.5° 58.4 klar Trübung deutlicher. 12. n. 0 0 0,004 9.9' 1.1° 44.2 „ Stark getrübt 13 0 0 0 10.2° 1,0° 50.7 „ 14 0 0 0,635 12.3° 2.9° 73.4 trüb Bei Carpinus Loch tiefer gebohrt 4. naclim. 15 0 0 2,279 18.6° 6.7° 77.5 Eegen 16 0 0 1,420 12.0° 3.5° 46.7 klar 17 0 0 0 6.1° 0.9° 55.7 „ 18 0 0 0,011 5.5° 0.1° 53.9 trüb 19 0 0 0,120 6.5° -1.3° 48.4 klar 20 0 0 0,070 7.9° -3.3° 51.6 " Alte Löcher geschlos- sen, neue löclier gebohrt 4. nachm. 21 0,470 0,042 0,389 10.5° -2.5° 54.7 trüb ffelit der Löcher- erneuerung schon deutlich 22 0,013 0,011 0,035 8.8° 0.3° 64.9 Uar 23 0,160 0,090 0,018 11.2° -0.5° 54.4 » 24 0 0 0 5.5° 1.4° 76.3 trüb 25 0,176 0,100 0 11.0° 3.8° 85.1 Hegen 26 10. vorm. 27 4. nachm. „ 0 1,175 1,310 0 0,627 0,380 0 3,410 9.3° 14.0° 6.3° 4.5° 80.7 71.3 trüb Bei Carpinus, altes Loch geschlossen, neues Loch gebohrt Summe 2,485 1,007 — 10. vorm. 28 4. nachm. „ 1,210 0,770 0,270 0,170 1,750 1,830 10.4° 6.8° 73.6 trüb Summe 1,980 0,440 3,580 4. nachm. 1. m 3,010 1,070 7,390 14.3° 7.0° 87.3 Eegen Der Saft bei Cornus I etwas getrübt 10. vorm. 2 4. nachm. „ Summe 2,400 0,870 3,270 0,990 0,340 1,330 10.360 5,000 15,360 17.3° 3.3° 64.0 trüb Effekt der Bohr- iocherneuerung deutlich 10. vorm. 3 4. nachm. „ 1,120 0,410 0,655 0,260 6,230 2,420 13.4° 0.1° 65.5 trüb Bei Carpinus Trü- bung durcli weiße Hefemassen Summe 1,530 0,915 8,650 10. vorm. 4 4. nachm. „ 0,610 0,390 0,490 0,240 0,590 1.070 13.8° 0.4° 58.0 klar Summe 1,000 0,730 1,660 4. nachm. 5 0,.590 0,490 0,640 13.3° 2.8° 55.0 " ÜBER DIE AUSFLUßMENGE DES BLUTUNGSSAFTES EINIGER BÄUME. Ausfluß menge in Litern Temperatur Relative Feuch- tigkeit 0 ' Witter- ung Zeit Cornus con- troversa I. Baum Cornus con- troversa n. Baum Carpinus \yedoenfih 1 Maxima C Minima C Bemerlaing 10. vorm. 6 4. nachm. „ 0,185 0,062 0,500 0,251 0,070 0,070 14.6° 4.5° 71.8 trüb Sivtume 0,247 0,751 0,140 10. vorm. 7 4. nachm. „ 0,160 0,070 0.255 0,090 0,865. 0,275 18.8° 6.4° 58.5 Eegen Löcher tiefer gebohrt Summe 0,230 0,345 1,140 10. vorm. 8 4. nachm. „ 4,530 2,025 2,350 0,880 11,280 4,550 13.6° 2.9° 57.2 klar Summe 6,555 3,230 15,830 10. vorm. 9 4. nachm. „ Summe 10. vorm. 10 4. nachm. „ 5,020 2,170 8,190 5,400 1,790 2,490 0,850 3,340 2,160 0,590 9,840 6,600 16,440 12,560 6,420 19.2= 11.0° 6.4° 4.2° 74.4 8L7 trüb Kffekt der Löcher- vertiefung ist deutlich Summe 7,190 2,750 18,980 10. vorm. 11 4, nachm. „ 5,340 1,690 1,450 0,410 11,380 3,610 4.4° 0.3° 90.0 Schnee Simime 7,030 1,860 14,990 4. nachm. 12 5,750 1.080 8,150 9.1° -2.6° 71.7 klar 10. vorm. 13 4. nachm. „ 4,450 1,630 1,130 0,390 4,980 4,240 18.6° 4.1° 83.7 trüb Summe 6,080 1,520 9,220 10. vorm. 14 4. nachm. „ 5,370 1,900 1,160 0,360 5,790 1,422 12.5° 7.4° 94.7 Summe 7,270 1,520 7,212 10. vorm. 15 4. nachm. „ 5,030 1,650 0,960 0,330 1,490 0,270 13.0° 4.7° 57.9 trüb Altes Lochlhei Car- pinus geschlossen, neues.Loch gebohrt Summe 6,680 1,290 1,760 10. vorm. 16 4. nachm. „ Summe 10. vorm. 17 4. nachm. „ 3,750 1,320 5,070 3,080 1,300 0,735 0,250 0,985 0,680 0,210 18,300 5,000 23,300 7,050 3,070 8.2° 5.2° 3.3° 0.5° 73.9 94.4 klar Schnee Eflfekt der Löcher- erneuerung ist deutlich Summe 4,380 0,890 10,120 10. vorm. 18 4. nachm. „ 3,820 1,370 0,640 0,220 7,670 4.800 16.1° 5.1° 73.9 trüb Summe 5,190 0,860 12,470 4. nachm. 19 5,450 0,800 12,500 8.8° 1.0° 69.6 Eegen 10. vorm. 20 4. nachm. „ 3,450 1,200 0,510 0,180 3,850 1,800 9.3° 0.1° 43.4 Mar Summe 4,650 0,690 5,650 ART. 1. — MANABU MIYOSHI Zeit 10. Yorm. 21 4. nachm. „ Summe 10. vorm. 22 4. nachm. „ Summe 10. Yorm. 23 4. nachm. „ Summe 10, vorm. 24 4. nachm. „ Summe 10. vorm. 25 4. nachm. „ Summe 10. vorm. 26 4. nachm. „ Summe 4. nachm. 27 4. nachm. 28 11. nachm. 28 10. vorm. 29 4. nachm. „ Summe 11. nachm. 29 lü. vorm. 30 4. nachm. „ Summe 11. nachm. 30 10. vorm. 31 4. nachm. „ Summe 4. nachm. 1 IV 4. nixchm. 2 4. nachm, 3 4. nixchm. 4 4. nachm. 5 4. nachm. 6 Ausflußmenge in Litern Temperatur Cornus con- troversa I. Baum Cornus L, con- IC^f^rpm^is trover sa Lgdnensis TL. Baum^ 3,100 1,220 4,320 2,850 1,050 3,900 2,930 1,000 3,930 2,980 1,120 4,100 3,110 0,950 4,060 1,980 0,760 2,740 0,360 0,170 1,450 1,190 0,950 0,530 0,360 0,150 0,060 0,020 0,425 0,160 0,585 0,405 0,140 0,545 0,410 0,150 0,560 0,430 0,140 0,570 0,390 0,130 0,520 0.375 0,125 0,500 2,790 2,230 0,265 0,100 0,085 0,085 0,050 0,030 0,010 0 0 1,125 3,100 4,225 0,5.50 1,950 2,500 1,650 1,250 2,900 2,950 0,880 3,830 1,350 0,530 1,880 0,970 0,240 1,210 0,440 0,370 11,270 17,850 11,150 40,270 7,500 11,050 8,640 27,190 7,650 12,050 4,920 24,620 7,950 2,750 1,280 0,430 0,020 0 Maxima C 9.5° 9.3° 9.6° 10.4° 12.1° 11.0° 17.0° 16.7° 18.6° 13.8° 19.4° 10.6° 12.5° 8.4° 17.6° 17.0° 19.5° Minima C Relative Feuch- tisrkeit 0.8° 1.8° 4.9° 3.1° 3.9° 2.8° 4.3° 6.7° 8.4° 6.9° 1.6° -1.1° 5.1° 6.9° 6.7° 8.0° 45.4 67.9 83.7 69.0 67.4 81.7 76.3 59.1 72.2 81.5 53.6 41.9 .57.0 87.0 57.5 50.1 44.2 Witter- uncr klar trüb Beeren klar trüb Begen trüb Bemerkune; Das Loch bei Car- plnus tiefer gebohrt Effekt der Lochver- tietung selir deut- lich klar ÜBER DIE AUSFLUßMENGE DES BLÜTUNGSSAETES EINIGER BÄUME. TAB. IL* AustlTißmenge in Litern Durch- schnittliche Tem- lieratnr C Relative Feuchtig- keit 0 -• Witterung Zeit Cormts C071- trowrsa Carpirms yedoensis Bemerkung 11. vorm. 7 m. 1914 2.4 2.0 17.9= 65.5 etwas geregnet 8 3.4 2.2 13.5° 45.5 klar 9 3.3 2.7 4.5' 89.7 Regen 10 4.4 2.4 5.2° 75.7 klar 11 3.2 3.9 4.7° 71.5 ,, 12 8.2 4.5 5..5° 95.0 Regen 13 9.2 5.7 11.7° 97.5 „ 14 10.4 7.1 7.2° 92.7 „ 9. vorm. 15 11.3 6.2 4.2-^ 71.0 klar 16 12.0 4.8 5.4° 57.5 „ 17 9.7 9.9 7.9° 64.6 „ 18 19 11.9 8.9 108 19.7 9.7° 6. .3° 82.2 49.4 etwas geregnet klar Gewöhnliches Maximtiui bei Cnrnus 20 10.7 20.6 8.5° 60.0 ,^ 21 9.4 20.7 11.1° 82.4 „ 22 11.3 30.2 12.2° 72.7 Regen 23 24 11.7 8.9 34.5 34.1 12.4° 12.9° 81.6 95.8 Regen (iewölinliclies Maximum bei Carpimis 4. nachm. 25 30.9 26.8 13.9° 85.6 „ 26 S.2 19.0 9.3° 88.7 27 28 7.7 7.2 14,3 6.2 4.0° 5.9° 74.2 61.0 etwas geschneit Idar 29 6.3 1.7 8.1° 65.5 „ 30 5.2 0.5 8.1° 89.6 Regen 31 4.7 1.4 11.2° 63.2 Mar 1. IV 4.3 0.2 13.2° 57.2 jj 2 4.1 0 9.7° 81.2 „ 3 2.9 0 4.7° 79.7 „ 4 0.4 0 1.1° 71.1 Schnee 5 0 0 4.6° 67.2 klar ♦Vorliegende Beobachtungen wunlen von meinem Assistenten Herrn Db. S. Hibino gemilcht. 8 AKT. 1. — ^MANABÜ MIYÜSHI : Die Zifferen in Tab. I zeigen den ganzen Verlauf der täglichen Ausflußmengen, die im letzteren Teile der Blutungsperiode viele Schwankungen erfahren. In der Tabelle sieht man, daß unsere Versuchsbäume in den Tagesstunden also, von 10 Uhr vormittags bis 4 Uhr nachmittags, verhältnismäßig stärker bluteten als in allen übrigen Stunden. Tab. II stellt die Ausflußmengen eines Abschnittes der Blutungsperiode dar, hier ist der Kurvenverlauf im Gegensatz zu demjenigen der Tab. I beinahe glatt. Auf diesen und anderen Punkte werde ich weiter unten zurückkommen. Um die äußeren Einflüsse auf die Größe des Ausflusses genauer zu konstatieren, muß man, wie frühere Forscher getan haben, nur mit Topfpflanzen experimentieren.^^ Bei unseren im Freien stehenden Pflanzen war es natürlich nicht möglich derartige Versuche anzustellen ; wir können jedoch unseren wiederholten Beobachtungen gemäß schließen, daß in der kälteren Zeit von Januar bis Anfang oder Mitte März die Temperatur von größter Bedeutung ist, in späteren wärmeren Perioden hingegen die Feuchtigkeit der Luft und des Bodens eine große RoUe spielt. Fig. 1, 2 und 3 veranschaulichen die eben erwähnten Ver- hältnisse. Wie man in Fig. 1 und 2 sieht, zeigen die Kurven der Ausflußmenge und Temperatur beinahe gleichartigen Verlauf, die Tage größerer oder geringerer Ausflußmenge faUen mit denjenigen höherer, resp. niederer Temperatur zusammen. In Fig. 3 ist der Einfluß der Feuchtigkeit auf die Ausflußmenge erkennbar, und zwar bei Carpinus deutlicher als bei Cornus, während der Effekt der Temperatur bei beiden Versuchspflanzen in dem Zeitabschnitte nicht mehr zu voller Geltung kommt. Es muß hier bemerkt werden, daß diese Verhältnisse wegen der Inkonstanz äußerer 1) Vergl. WiELEK, Das Blaten der Pflanzen. (Cohns Beitr. z. Biol. d. Pfl. Bd. TL. 1893. p. 48 \\. 57), und aiicli Pfeffek, Pflanzenpliysiologie. H. Aufl. Bd. I. 1897. p. 245. üßEK DIE AUSFT.UßMENGE DES BLUTUNGSSAFTES EINIGER BAUME. Fig. 1. o°c na rsi r . - ^ / ßO 10^ '^4— H- >^ / ^^X- t \ ' 7>'' .-. — y "1 r ^< — iUy6 ö^- Jj^d. ^~.Z _i5.__ _,__ ^ ^ t' o ^^ 2--r 1 __ •-CS CS 10 AET. 1. — MANABU MIYOSHI Bedingungen nicht immer zu sehen sind, und ferner, daß ein spezifischer, ja sogar individueller Unterschied in dieser Beziehung zu existieren scheint. Auch dient Tal). III als ein eklatantes Beispiel des Feuchtig- keitseffektes auf die Ausflußmenge. Hier war die Bohrlocher- neuerung, die allerdings spät ausgeführt worden ist, wirkungslos geblieben, erst nach einem Regenguß, der bald darauf folgte, trat ein lebhafter Ausfluß ein. TAB. III. Zeit Ausflußmenge in Litern Carpbvus yedoensis Durchsohnitt- liche Temperatur C Kelative Feuchtigkeit Wittemng Bemerkung 2. IY.1913 3.0 12.8^ 56.4 Idar 3 1.8 12.0^ 70.7 ,. 4 1.6 12.8° 70.0 5 0.2 11.3° 68.6 6 0 16.6° 67.3 ,, 7 0 16.3° 71.3 „ Ein neues Loch wurde gebohrt. Kein SaftausfluE 8 — 12.3° 92.5 frühmorgens Eegen Durch Regen begann der SaftausfluE 9 10.8 5.9° 80.5 Hegen GroEer SaftausfluE 10 o.e. 7.3° 57.9 klar Zwischen 1-2 Uhr naclimittags hörte der SaftausfluE auf Wie in den oben stehenden Tabellen, besonders in Tabelle II ersichtlich ist, nimmt der Ausfluß an den späteren Tagen der Blutungszeit rasch ab, bis er beinahe sistiert. Dies ist aber kei- neswegs das Zeichen des totalen Aufhörens des Blutens. Denn ein lebhafter, oft erstaunlich großer Saftausfluß wird sogleich erzielt, sobald man der alten Erfahrung gemäß das Loch noch tiefer bohrt, oder den inneren Flächenteil wegschneidet, ev. ein neues Loch aufmacht. ÜBER DIE AUSFLUßMENGE DES BLUTUNGSSAFTES EINIGER BÄUME. 11 Als Beispiele mögen folgende Fälle bei Carpinus yedoensis aus Tab. I. Erwähnung finden. I. Fall. Am Mittag des 14. Februar wurden 0,635 1 Saft für die letzten 24 Stunden gefunden. Sofort wurde das Loch tiefer gebohrt, worauf die Saftmenge bis 4 Uhr nachmittags des nächsten Tages auf 2,278 1 stieg. Darauf nahm, der Ausfluß bald ab. IT. Fall. Vom 19. mittags bis nächsten Mittag war die Saft- menge 0,070 1. Sodann wurde das Loch geschlossen und ein neues Loch gebohrt. In den folgenden 28 Stunden erreichte die Aus- flußmenge 0,389 1. Am Tage darauf floß aber nur wenig Saft aus. III. Fall. L^m 4 Uhr nachmittags des 27. war die Ausfluß- menge der vorangehenden 28 Stunden 3,410 1. Es wurde sogleich ein neues Loch gebohrt, aus welchem in den nachfolgenden drei Tagen 3,580, 7,390, resp. 15,300 1 Saft pro 24 Stunden flössen. Dann nahm der Ausfluß rasch ab. IV. Fall. Um 4 Uhr nachmittags des 15. März wurden 1,760 1 Saft für die vorangehenden 24 Stunden gefunden. Ein neues Loch wurde sogleich gebohrt und die Ausflußmenge erreichte am nächsten Mittag 23,300 1. Nachher trat eine schnelle Ver- minderung ein. V. FaU, Von 4 Uhr nachmittags des 27. bis zu derselben Zeit des nächsten Tages kamen aus dem Bohrloch 0,370 1 Saft. Darauf wurde das Loch tiefer gebohrt und in den nachfolgenden drei Tagen flössen 40,270, 27,190 resp. 24,620 1 Saft pro 24 Stunden aus. Nachher wurde die Ausflußmenge schnell geringer. Fig. 4 zeigt auch einen anderen Fall bei Carpinus sehr deut- lich. Der Saftausfluß hörte schon vom 4. April bis zum Morgen des 8. fast auf. Sodann wurde gegen Mittag ein neues Loch gebohrt, aus welchem 22 1 Saft bis nächsten Mittag ausflössen. Obgleich der Ausfluß am folgenden Tage etwas abnahm, erreichte er am 12 ART. 1. M ANABU MIYOSHI Fig. 4. o^ ___ __ _ -- - 1 _l iL 7 on OR _ .. . ...... Lu. -^ ^U Aü tt /'^ T4 / ^5^4^ ^^ 2 -^^ J^^A IC 04 4L -4: Z^^^__^ J JI _ _ _ ?l^ ^ = t 4^ '. J. SoLLis, On Onychaster, a Carboniferous Brit tie- Sta r ; Philos. Transact, Ser. B, Vol. 204, 1913;— Th. Mobtensen, On the Alleged Primitive Ophinroid, Ophioteresis elegnns Bell, &c. ; Mindeskr. for .Tapetxjs Steenstrup, 1913. 4 ART. 2. — H. MATSUMOTO : h. Those which are divided into lialves by a series of small pores or by a moniliform pore, are found in those forms in which the dorsal side of the arms is entirely protected. E. g. Microphiura ; Ophiologimus ; Ophlothamniis emend. ; " Ophiactis " pars, i.e. my Ampliiadis ; " Ophioconis " pars, i.e. my Ophiuroconis and Ophiurodon ; Astroph'nira ; àa, 4. Those forms with quadrangular and stout teeth have oral frames with well developed lateral wings for the attachment of voluminous masticatory muscles. E. g. Amphiuridce mihi ; Opino- trichidœ ; Ophioceramîs ; Ojjhionereidinœ mihi ; and Ophiocomidœ. B. Systematic. 1. Astrocevas, Trichaster and Eurycde have a certain common structure, by which they may be distinguished from either Asteronyx or Aster oschema. 2. Astrotoma and its allies have a certain distinctive character in contrast to Asteroporpci, Astrochele, Gorgonocephahis, Astroden- drum, Astnocladus and their allies. 3. The Amphiuridce mihi and Ophiotrichidœ stand in a very intimate relation in ther internal structures. 4. " Ophiactis " pars, i.e. my Amphiactis is a connecting link between the Ophiaccmthidœ and Amphiuridce. 5. The Ophiolepididce, Ophiodermatidce and Ophiocomidœ form together another compact group. 6. The " Amphiuridce " pars, i.e. my Ophiochitonidce are by no means allied to the genuine Amphiuridce, but are very near to the Ophiodermrdidce and Ophiocomidce. 7. Ophiopsila is, as a matter of fact, notliing else than an ally of the Ophiocomidce. 8. " Ophioconis " pars, i.e. my OphiuTOConis and Ophiurodon, MONOGRArH OF JAPANESE OPHIUROIDEA. 5 and " Ophioclueta " pars, i.e. my Ophkirochceta, may be clearly distinguished from Ophiolimna emend, by certain internal struc- tures. Prefixing so much, I now proceed to the exposition of my own views, leaving them to be judged by their own merits. I have several representatives of Palœozoic ophiurans in hand, and I purpose to publish a revision and classification of them in the near future. The greater part of the present study was done in the Zoological Institute of the Imperial University of Tokyo, to which the type specimens of all the new species described in this paper belong. Such new genera as Astrothanmus, Ophiosemnotes, Ilaplo- pJiiura, Amphiopliiura, Ophiosonella, Ophiuroconis, OjjJiiurochceta, &g,, are directly or indirectly based upon the specimens in Dr. H. L. Clark's charge. All the text and plate figures in this paper were drawn by me. Subclass I. ŒGOPHIUROIDA Matsumoto. Ophiuroidea with external ambulacral grooves and without ventral arm plates. Radial shields, genital plates and scales, oral shields, peristomial plates and dorsal arm plates also absent. Ambulacral plates alternate or opposite ; in the latter case, they may often be soldered in pairs to form the vertebrae. Adambula- cral plates, i.e. lateral arm plates, subventral in position. Madre- porite either dorsal or ventral, often large and similar in shape to that of an asteroid. This subclass mostly consists of Palaeozoic forms, and lacks all the fundamental characters by which the recent ophiurans are clearly distinguished from the asteroids. Indeed, the distinction of 6 ART. 2. — H. MATSUMOTO : the present subclass from the asteroids depends merely upon the different development of certain common structures. Subclass II. MYOPHIUROIDA Matsumoto. Opliiuroidea without external ambulacral grooves, and with ventral arm plates. Eadial shields, genital plates and scales, oral shields, peristomial plates and dorsal arm plates usually present ; but sometimes, some or all of them may be rudimentary or absent. Ambulacral plates opposite, usually completely soldered in pairs to form the vertebra?. Madreporite represented by one, or sometimes all, of the oral shields. This subclass includes certain Palaeozoic forms and all the ophiurans from the Mesozoic downwards. The Palaeozoic Myophiuroida appear to me to represent a dis- tinct order by themselves ; but I defer the question to a future paper. Order i. phrynophiurida Matsumoto. Disk and arms covered by skin. Eadial shields either very rudimentary or long and bar-like. The radial shield and genital plate of either side of each radius articulate with each other by means of a simple articular face or a transverse ridge of either plate. Peristomial plates large, entire, double or triple. Oral frames entire, without well developed lateral wings. Lateral arm plates subventral or ventral in position, while the dorsal arm plates are either entirely absent or rudimentary, so that the dorsal side of the arms is mostly unprotected. Vertebral articulation zygospondyline or streptospondyline. MONOGRAPH OF JAPANESF- OPHIUKOIDEA. 7 Key to families of Phvijnophiurida. A — Kadial shields very rudimentaiy or more or less long und bar-like ; arms not exceedingly long, always simple ; lateral arm plates sub- ventral in position ; arm spines not confined to the ventral side of the arm ; vertebrae rather slender, not exceedingly short and dis- coidal ; vertebral articulation either zygospondyline or streptospon- dyline ; both upper and lower muscular fossae of the vertebrae large OpTdomyxidœ. A A — Kadial shields very long and bar-like ; arms exceedingly long, either simple or branched ; lateral arm plates ventral in position ; arm spines confined to the ventral side, all serving as tentacle scales ; vertebrae very stout, exceedingly short and discoidal ; vertebral arti- culation typically streptospondyline ; upper muscular fossae of the vertebrae extremely large, the lower very small. a — Teeth very stout, arranged in a single vertical row ; oral and dental papiUae not very well developed ; adorai shields very stout ; arms not annulated by rows of hook-bearing granules ; arm spines covered by thick skin Trichastendœ. aa — Teeth and dental papillae, often also oral papillae, all similar, spiniform ; adorai shields not very stout ; arms annulated by double rows of hook-bearing granules ; arm spines naked or covered by very thin skin GorgonocepJiàlidœ. Family 1. Ophiomyxidae (Ljungman, 1866) mihi, 1915. Disk and arms covered by a soft skin. Radial shields very rudimentary or more or less long and bar-like. The radial shield and genital plate articulate with each other by a simple face. Genital scales slender, bar-like, articulated with the genital plate at a distance inw^ards from the outer end of the latter. Peristomial plates very large, either double or triple, and thin or entire and 8 AET. 2. — H. MATSUMOTO : very thick. Oral frames either long and slender or short and very stout, without well developed lateral wings. Arms slender, not exceedingly long, always simple. Dorsal arm plates absent or rudimentary, lateral arm plates subventral in position, so that the dorsal side of the arms is mostly unprotected. Arm spines not confined to the ventral side, skin-covered. Vertebrsß rather slender, not very short and discoidal, often incompletely calcified and more or less divided into halves by a longitudinal, fusiform pore. Vertebral articulation zygospondyline, the articular peg being well developed, or streptospondyline, the peg being very rudimentary or entirely absent. Both upper and lower muscular fossae large. This family includes twenty-two genera, which may be grouped into two subfamilies as follows : Subfamily 1. Ophiomyxince mihi, 1915. — Oral sliields small, separated from the first lateral arm plates by the outer lobes of the long and slender adorai shields ; vertebra long and very slender, often divided into halves ; wings of the vertebrae very much thinner laterally than dorsally, so that the vertebrae are fairly rhomboidal in dorsal view ; vertebral articulation zygospon- dyline, the articular peg being well developed. I. Second oral tentacle pores opening outside the oral slits. a. Teeth present, arranged in a single vertical row. Ophiohelus Lyman, 1880. Ophiosciasma Lyman, 1878. Ophiogeron Lyman, 1878. Astro ger on Verrill, 1899. Ophlostyracium Clark, 1911. Opliiooynodus Clark, 1911. h. Teeth almost absent. Ophiosyzygus Clark, 1911. MONOGRAPH OF JAPANESE OPHIÜROIDEA. 9 Oph'ioleptoplax Claek, 1911. II. Second oral tentacle pores opening entirely within the oral slits. a. Teeth and oral papillae acute, not widened and serrate. Ophioscolex Müller à Tkoschel, 1842. Neoplax Bell, 1884. Ophiostlba Matsumoto, 1915. Ophiomora Kœhler, 1907. b: Teeth and oral papillae widened and serrate along the fi'ee end : Ophiohyalus Matsibioto, 1915. Opliiomyxa Müllek & Teoschel, 1842. Ophiodera Yerrill, 1899. Ophioliymen Clark, 1911. Subfamily 2. Ophiohyrsince Matsumoto, 1915. — Oral and adorai shields very thick, intimately joined to each other, the latter being entirely proximal to the former ; vertebrae short and stout, always entire ; wings of the vertebrae very thick laterally as well as dorsally ; vertebral articulation streptospondyline, the arti- cular peg being very rudimentary or entirely absent. I. Oral shields absent, except in one interradius ; a single genital slit to each interradius. Ophioschita Clark, 1911. IL Oral shields well developed ; two genital slits to each interradius. Ophiophrixus Clark, 1911. Ophiobyrsa Lyiman, 1878. Ophiobyrsella Verrill, 1899. Ophiosm'ilax Matsumoto, 1915. OpMohrachion Lyman, 1883. ÏO ART. 2. — H. MATSUMOTO : The difference between the Ophiomyxinœ and Ophiohyrsinœ appears to me to be very sharp, especially in the internal struc- tures. In Ophiodera anisacantha, the peristomial plates are triple, the unpaired one filling up, Hke a wedge, the outer open angle formed by the paired ones, wliich are very long and boot-shaped. The oral frames are stout, with a distinct groove for the ambula- cral ring canal. The first few vertebrae are very short, with very thin wings. The genital plates are very slender, articulating with the genital scale at a distance inwards from the outer end. The articulation of the genital plate and radial shield is very simple, without any articular condyle or pit. Each vertebra outside of the base is composed of a slender body and very thin wings. The vertebral articulation is zygospondyline. The articular peg is well developed, connected with the halves of the articular shoulder by a short, narrow ridge, so that the whole trio is shaped like an M, of which the upper open angle corresponds to the pit for the articular umbo and the lower open angles to the pits for the articular knobs of the next vertebra. The articular umbo is very stout, elongated rhomboidal, while the articular knobs are rather feeble, separated fi'om each other by the pit for the articular peg. In Ophiomijoca flacclda Müller & Teoschel and 0. austrcdis LüTKEN, the peristomial plates are double, nearly soldered together ; especially in very small specimens of the last species, the double plates are very firmly soldered together, so as to appear almost entire. The vertebral articulation of the last species is rather peculiar : the trio of the articular peg and halves of the articular shoulder are closely set side by side, so as to appear like a united mass leaving a furrow-like pit for the articular umbo of the next vertebra above them ; the articular umbo is very prominent, wMle the articular knobs are scarcely developed, being represented only MONOGRAPH OF JAPANESE OPHIUKOIDEA. 11 by two insignificant ridges between the pits for the articular peg and the halves of the articular shoulder. In Ophiohyalus gotoi the peristomial plates are usually double, with oblong halves, but are triple in some interradii, the unpaired one being quad- rangular and overlapping the paired ones. In Ophiostiha liidekii the peristomial plates are double, often with unequal halves, one of which overlaps the other. According to Lyivian, Ophioscolex glacialis Müller & Teoschel has triple peristomial plates and slender, cylindrical oral frames, without distal wings.'-* In Opliiohyrsa rudis Litnian, the internal structures of which were studied by Lyman, the characteristics of the Ophiohyrs'mce appear to be well realised. The whole oral skeleton is very com- pact. The peristomial plates are entire, stout, intimately soldered to the oral frames, which are also stout. The radial shields, as well as the genital plates and scales, are comparatively small, the plates and scales bounding only the outer part of the long genital slits. The vertebrae are stout, short, more or less discoidal, the wings being very thick laterally as well as dorsally ; a number of basal vertebr«3 are especially discoidal, like those of the next two families. The vertebral articulation is perfectly streptospondyline, the articular peg being entirely absent. The articular shoulder is divided into two long, stout, parallel condyles by a median groove. The articular umbo is very stout, the downward prolongation con- nected without any boundary with the upward prolongation of the soldered mass of the articular knobs, so that the whole of the umbo and knobs taken together is long dumb-bell shaped. In Ophiosmilax mirabilis the vertebral articulation is essentially similar, but there persists a well marked line of contact between the articular umbo and the soldered mass of the articular knobs. 1) This peculiarity will be discussed in a future paper. 12 ABT. 2. — H. MATSUMOTO : In both species, the upper and lower mnscular fossœ are both large, a characteristic of the " common oploim-ans " in contrast to the next two famihes, to which the Ophiohyrslnœ rather approach in almost all characters. Key to Japanese genera of Ophiomyxinœ, A — Second oral tentacle pores opening outside the oral slits. a — Disk free of spines ; radial shields absent ; arm spines not very consjpicuonsly thorny ; a few oral and dental papillae present ; no teeth. h — Dorsal arm plates entirely absent ; arm spines connected together by a web-like membrane OpMosyzygus. hb — Rudimentary dorsal arm x^l^tes present, entire and hyaline ; arm spines free from each other OpMoleptoplax. aa — Disk beset with a number of long, acute, conspicuous spines ; radial shields conspicuous, each bearing one long, acute, conspicuous spine ; arm spines very cons]3icuously thorny ; a few spiniform oral papillae, besides one short and rounded outermost one, which arises from the fii'st ventral arm plate OpMostyracium. AA — Second oral tentacle pores opening within the oral slits. c — Teeth and oral papillge triangular, acute, not hyaline ; radial shields very rudimentary ; marginal disk scales present ; dorsal arm plates entirely absent ; arm spines converted into compound hooks Ophiostiba. cc — Teeth and oral papillae flattened, with widened and serrate end more or less hyaline. d — Marginal disk scales present. e — Rudimentary dorsal arm plates always entire ; radial shields very rudimentaiy ; vertebrae divided into halves, except in several basal arm joints ; arm spines converted into compound hooks . . Ophiohyalus. ee — Rudimentary dorsal arm plates divided into several secondary plates, except in the distal arm joints ; radial shields not very rudimentary ; ÄIONOGRAPH OF JAPANESE OPHIUROIDEA. 13 vertebrae entire, except in very distal arm joints ; arm spines not converted into compound hooks, except in the very distal arm joints Ophiomyœa. dd — Marginal disk scales absent. f- — Radial shields present ; dorsal arm plates usually absent (or some- times present, being divided into several secondary plates, as in the Japanese species) ; arm spines free from each other. . OpModera. ff — Radial sliields absent ; dorsal arm plates entirely wanting ; arm spines connected together by a web-like membrane .... OpMohyTnen. OpJiiosy^ygns disacanthus Clark. OpMosyzygus disacanthus : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 276, fig. 139. Off Kagosliima Gulf; 103-152 fathoms (Claek). Uraga Chan- nel ; 88 fathoms (Clakk). Ophioleptoplax megaxyora Clark. Op7iioIepfo2:)lax megapora : Ctj^rk, Bull. U. S. Nat. Mus., LXXV, 1911, p. 279, fig. 141. Eastern Sea; 71 fathoms (Clark). Ophiostyraciuni tvachyacanthiini Clark. OpMosfyracivm trachyacanthum : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 253, fig. 124. Eastern Sea ; 440 fathoms (Claek). Ophiostiba Matsumoto, 1915. Disk and arras covered by a soft, thin skin, which contains some granules in the former. Marginal disk scales present. Radial shields very rudimentary, forming a continuous row with the marginal disk scales. Teeth and oral papillae present, triangular, 14 ART. 2. — H. MATSUMOTO : with acute tips. Dental papillsG absent. Second oral tentacle pores opening entirely within the oral slits. Yertebrse more or less divided into halves by a longitudinal fusiform pore in the outer half of the arm. Dorsal arm plates entirely absent, lateral arm plates more or less subventral in position, so that the dorsal side of the arm is mostly unprotected. Two or three arm spines, all converted into compound hooks. Tentacle scales absent. This genus differs from Ophloscolex cliiefly in the presence of the marginal disk scales and in the conversion of the arm spines into compound hooks ; and from Neoplax in the fewer arm spines, wliicli are all converted into compound hooks, and in the absence of the tentacle scales. OpJiiostiba hidekii Matsumoto. Opliiostiba hideJcii : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, ILXyn, 1915, p. 47. Two specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Diameter of disk 3.5 mm. Length of the longest arm 1(3 mm. Width of the same at base 0.8 mm. Disk hexagonal, with concave interbrachial borders, ratlier convex dorsally, covered by a soft skin, wliich contains fine, scattered, calcareous granules. Marginal disk scales present, but rudimentary, liable to be overlooked. Eadial shields very rudi- mentary, insignificant, forming a continuous row with the marginal disk scales. Genital slits very small and short. Oral shields rhomboidal, with perfectly rounded outer and lateral angles ; convex, slightly longer than wide ; all serving as madreporites. Adorai shields narrow, but with widened outer end, wliich entirely sejoarates • the oral shield from the first lateral arm plate ; adradially concave, meeting each other within. Four or MONOGRArn OF JAPANESE OPHIUROIDEA. 15 O c i^ „. ... , , ,. , .^ towards the extremity of the of two arm joints somewhat near the extremity. «^ xi6. /. Dorsal view of the vertebra of the arm. Lateral arm plates slight - fourth free arm joint, x 14. g. Dorsal view of the vertebra of the tenth free arm joint. X U. ly flarcd, SUCCCSSlve plateS not in contact with each other, but separated by a naked space, which is widened up- wards and continued into a large naked space bounded by the dorsal and lateral arm plates and the vertebra. First ventral arm plate not very small, quadrangular, with strongly curved outer edge, much wider without than within. Those beyond nearly rhomboidal in outline, with a conspicuous notch at the outer end and a half pore for the tentacle at each lateral angle ; much longer than wide, being widest at the outer ends of the tentacle pores ; successive plates not in contact with each other, MONOGRAPH OF JAPANESE OPHIUKOIDEA. 19 except in the disk. The lateral arm plates do not however meet each other in the ventral median line, so that there is left here a naked, depressed space, which is especially well marked near the extremity of the arm. Two arm spines, subventral, miequal, glassy, all converted into compound hooks, with a series of hook- lets along their ventral side, covered by a thin, transparent mem- brane ; the lower one is much larger than the other. In some basal joints, there occurs on the lateral arm plates one more spine, which is placed on the dorsal margin of the plate and also bears a series of booklets on one side ; it is larger than the other two, and nearly as long as the corresponding arm joint. No tentacle scale. Colour in alcohol yellowish white. Ophiomyxa aiistralis Lütken. OpJdomyxa austrcdis : Lütken, Addit. ad Hist. Oph., IH, 1869, p. 45 & 99 ; Lyman, Kep. Challenger, V, 1882, p. 246 ; Lyman, Bull. Mus. Comp. Zool., X, 1883, p. 274; Brock, Zeitschr. wiss. ZooL, XLVII, 1888, p. 532; Kœhler, BuU. Sei. Fr. Belg., XXXI, 1898, p. 123; Kœhler, Ibid., XLI, 1907, p. 341 ; Clark, Mem. Austral. Mus., IV, 1909, p. 547 ; Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 271. Four specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Oflf Suno Saki, Sagami Sea ; 41-50 fathoms (Clakk). Korea Strait ; 59 fathoms (Claek). Eastern Sea ; 95 fathoms (Clark). Amboina ; 100 fathoms (Lïman). Sulu Sea ; 82-102 fathoms (Ly^ian). Tongatabu Is. ; 18-240 fathoms (Lyman). Near Fiji Is. ; 310-315 fathoms (LYikiAN). Southern Australia and Bass Strait; 38-120 fathoms (Lütken, Lyman, Claek). New Zealand (Kœhlee). My specimens are all small, the largest being 7 mm. across 20 ART. 2. — H. MATSUMOTO : 9%^M*-. the disk. The colour is i3urpUsh brown in alcohol, the disk being beantifnlly spotted with darker shades. The rudimentary dor- sal arm plates are entire and transparent at the outer parts of the arms, but are divided into two or three opaque, secondary plates in the basal arm joints. The smallest speci- men, with the disk diam- eter of 3.5 mm., has entire dorsal arm plates throughout, and has only one or two granule-like oral papilloD on either side and less flattened and not serrate teeth, besides a Fig. 3. OphlomyX't audraUs. a. From above. x7. h. From below. x7. c Side view of four arm joints t i i ne^r disk. x7. d. Dorsal view of two arm joints at VCry short aud Wide dlstal the middle part, x 7. c. Young specimen ; from below. xl4. within the oral slit. The papilla, situated deep second oral tentacle pores of the same specimen open outside the oral slits, and the first ventral arm plates are comparatively large. A specimen with the disk diameter of 4 mm. has three of the jaws with typical papillation, and two papillated as in the smallest specimen mentioned. Probably, the oral papilla3 next the apex of the jaws are the first to appear, and the more distal ones are formed later. The distal papilla of the smahest specimen is not a true oral papilla, lying at a higher level than the ordinary ones. It may be distinguished as the primary oral papilla in contrast to MONOGRAPH OF JAPANESE OPHIUEOIDEA. 21 the genuine or secondary oral papilla?. The primary oral papilla) are usually predominant in many genera, which have exposed second oral tentacle pores. Ophiodera anisacantha (Clark). OpMomyxa anisacantha : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 277, fig. 136. One specimen; Suruga Gulf; 115 fathoms. One specimen Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Five specimens ; Sagami Sea. One specimen ; Uraga Channel ; 125 fathoms. Uraga Channel; 197 fathoms (Claek). Suruga Gulf; 108-131 fathoms (Claek). Off Suno Saki, Sagami Sea ; 83-158 fathoms (Claek). Eastern Sea ; 95-181 fathoms (Claek). Diameter of disk 28 mm. Length of arms 140 mm. Widtli of arms at base 4 mm. Disk pentagonal or five- lobed, with concave or slightly notched interbrachial borders, covered by a thick skin, which is finely and almost concentric- ally wrinkled in alcohol. The radial shields can be observed by drying the speci- mens ; they are exceedingly small and club-shaped. Genital slits long, not quite reaching the disk margin, narrow in the outer part, bordered adradially by the stout genital plates, much widened in the inner part Fig. 4. OpUodera anisacantha. x3. a. From above. 5. From below, c. Side view of four arm joints near disk. 22 AET. 2. — H. MATSUMOTO .* lying within the genital plates, where they present elongated oval outhnes. The genital plates and scales are distinguishable, though they are covered by the skin. The oral skeleton can be observed by drying the specimens. The oral shields are transversely oval, or rather rhomboidal, with an obtuse inner angle, and wddely rounded outer and lateral angles. Adorai shields small, curved abradially as a whole, tapering in- wards, where they do not meet each other. Four oral papillae on either side, flat, thin, translucent, and finely serrate along the free edge. Five to seven teeth, short, flat, with rounded, finely serrate or nearly entire end ; the uppermost one is longer and more or less conical. Below the first oral tentacle pore in the oral slit, occur one to three spiniform papilla). Arms slender, covered by a thick skin. The arm plates and vertebra3 can be observed by drying the specimens or by boiling an arm piece in potash. The vertebrae are rhomboidal when view- ed from above, much wider than long, with a conspicuous median groove. Between them, there are, in the proximal parts of the arms, a few irregular scales, perhaps representing the dorsal arm plates. Lateral arm plates subventral, meeting below, where they are soldered together and with the ventral arm plates of the corresponding joints. The ventral arm plates are eight- sided ; two sides embracing a conspicuous notch at the outer end of the plate ; inner sides the longest, straight, meeting each other in a wide angle ; inner lateral sides concave, corresponding to the tentacle pores ; wider than long, widest at the outer ends of the tentacle pores. Three or four arm spines, short, stout, conical, blunt, enclosed in skin, rough at the end ; the uppermost one is the longest, nearly as long as the corresponding arm joint, somevfhat isolated from the rest which are near together. The arm spines MONOGEAPH OF JAPANESE OPHIUEOIDEA. 23 are smaller in the basal joints, and only one is present in tlie first and second, but two in the third and fourth. No tentacle scale. Colour in alcohol : disk bluish gray ; arms light yellow ; disk and arms of younger specimens light grayish pink. In the smaller specimens, the disk skin is seen to contain, when examined under a microscope, very fine scales scattered in it, much resembling the perforated spicules of a holothurian ; but in the larger ones they are entirely absent. The oral papilla?, the papilla? just below the first oral tentacle pores, the teeth and the arm spines appear to increase in number witli the growth of the animal. In tlie specimens before me, the arm spines are not so long and so slender as are shown in Claek's figure. Further, his figure appears to me to be inaccurate in so far as it represents the arm spines as being two in number and equally long already in the first arm joint. Ophiohynien gymnodisciis Clark. OpMoJiymen gijmnodiscus : ClaPlK, Bull. U. S. Nat. Mns., LXXV, 1911, p. 277, fig. 140. Eastern Sea ; 107-139 fathoms (Clark). Key to Japanese genera of OpMohyrsinœ, A — Radial shields more or less long and bar-like ; disk or, at least, radial shields spinulate ; arm spines spiniform, not all converted into compound hooks. a — Dorsal arm plates entirely absent ; arm spines very long and serrate. , OpMophrixus. aa — Rudimentary dorsal arm plates present, being divided into a num- ber of secondary plates ; arm spines rather short and not serrate. . . Ophiobyrsa. 24 . ART. 2. — H. MATSUMOTO : A A — Eadial shields very rudimentary, disk entirely free of spines ; arm spines all converted into compound hooks OpMosmildx. Oiyhiophrixus acanthinus Clark. OpJiiophriœus acanthinus : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 273, fig. 137. Off Seno Umi, Suruga Gulf; 94-150 fathoms (Clakk). Key to Japanese species of Ophiohyrsa, A — Each dorsal arm plate represented by a zone of a mosaic of small, iri'egular, secondary plates, each of which carries a spinule ; six or seven arm spines not webbed together, the lowest one is converted into a compound hook and serves as a tentacle scale acantJiinohracIiia. AA — Each dorsal arm plate represented by a row of secondary plates, of which the median two are the largest, none of them spinulate ; four arm spines connected together by a web-like membrane synaptacantJm. Ophiohyrsa acanthinobracJiia Clark. OpJiiohyrsa acanthinobracliia : Clark, Bull. U. S. Nat. Mus., LXXV, 1911. p. 269, fig. 134. Suruga Gulf; 108 fathoms (Claek). Eastern Sea; 95 fathoms (Claek). Ojjhiohyrsa synaptacantJia Clark. OpJdohyrsa synaptacantlia : Clark, loc. cit., p. 270, fig. 135. Eastern Sea ; 152 fathoms (Claek). Ophiosmilax Matsumoto, 1915. Disk and arms covered by a thick skin, which may contain MONOGEAPH OF JAP.\NESE OPHIUEOIDEA. 25 some fine scales in the former. Radial shields very rudimentary (apparently " absent "), lying marginally. Oral sKts small. Single oral papiUa on either side, and two or three dental papilla) at the apex of each jaw. Teeth in a single vertical row. Teeth and papillœ all alike, stout, stumpy, conspicuously thorny at tip. Second oral tentacle pores opening outside the oral slits, each provided with a thorny, stumpy papiUa, which arises from the adorai shield. Arms verticaHy coiled. Vertebra very stout, with saddle-shaped articulation. Dorsal arm plates entirely absent, while the lateral arm plates are subventral, so that the dorsal side of the arms is covered merely by a naked skin. Two or three arm spines, all converted into compound hooks. Tentacle scales absent. This genus rather resembles OpUophrixus in the total absence of the dorsal arm plates, but differs from it in the very rudimentary radial shields, in the peculiarities of the teeth and papilla?, and in the conversion of all the arm spines into compound hooks. This last character is also found in Ophiobrachlon, but the present genus differs from it in the total absence of the disk spines, in the peculiarities of the teeth and papilla?, and in the fewer arm spines. Oj>7iiosmilax inirahilis Matsumoto. OpUosmilax inirahilis : Matsümoto, Proc. Acad. Nat. Sei. Philadelphia, LXYn, 1915, p. 50. One specimen ; Mora- out- Oisegaké, Sagami Sea ; 300 fathoms. Diameter of disk 2 mm. Length of arms 12 mm. Width of arms at base 0.8 mm. Disk flat, pentagonal, with concave interbracliial borders, covered by a thick skin, which contains very fine, tliin, trans- 26 AKT. 2. — H. MATSUMOTO parent scales. Radial shields very rudimentary, insignificant, lying on the disk margin. Genital slit short and very small. Oral shields trian- ^>. gular, with convex outer border. Adorai shields Eig. 3. OxMosmilax mircibiUs. x20. a. From above. h. From below, c. Side view of three arm joints near disk. large, quadrangular, longer than wide, wilder outwards than inwards, fully meet- ing with each other within. Oral slits short, fairly closed up. Single oral papilla on either side, stout, stumpy, conspicu- ously thorny at tip, turn- ed up ventrally, instead of projecting towards the oral slit. Two or three dental papillae at the apex of each jaw, similar in shape and size to the oral papilla, also turned up ventrally. Teeth in a single vertical row, also stout, stumpy and thorny at tip. Second oral tentacle pores opening entirely outside the oral shts, each provided with a stumpy and thorny papilla, which arises fi'om the adorai sliield. Arms more or less vertically coiled, covered by a thick skin. Dorsal arm plates entirely absent. Lateral arm plates subventral, strongly flared laterally. First ventral arm plate large, quadran- gular, with rounded angles, slightly longer than wide, much wider without than within. Those beyond larger, hexagonal, the. inner and inner lateral sides being very short ; outer angles per- fectly rounded ; as long as, or slightly longer than wide, feebly MONOGRAPH OF JAPANESE OPHIUEOIDEA. 27 calcified and transparent, except the outer and lateral peripheries, where the calcification is complete and the plate is opaque. Two or three arm spines lying flat on the arm, all coverted into com- pound hooks, transparent, except the ball-like basal portions, wliich are opaque. The uppermost two are subequal, about two- thirds as long as the corresponding arm joint, wliile the third, the lowest, is very small and about half as long as the same. The smaller compound hooks have two or three denticles in one plane, while the larger ones have six or seven denticles in two planes, making certain acute angles with the axis of the spine. Tentacle pores small, free of tentacle scales. Colour in alcohol brownish yellow. Family 2. Trichasteridse (Dodeelein, 1911) mihi, 1915. Disk and arms covered by a thick skin, which may contain fine scales or granules. Eadial shields long, bar-like, forming a system of radiate ribs. The radial shield and genital plate arti- culate with each other by means of a transverse ridge on both the plates. Genital plates long and stout, while the genital scales are rather very small. Oral shields small, adorai shields very stout. Teeth very stout, arranged in a single vertical row ; oral and dental papillaB rudimentary as a rule, sometimes absent. Peristomial plates entire, stout, soldered with the also very stout oral frames. Arms exceedingly long, simple or dichotomously branched, vertically coiled. Dorsal arm plates either entirely absent or represented by very insignificant double rows of fine sub- cutaneous scales, which are soldered together to form two, right and left, secondary plates in the free arm bases of the adult in certain genera ; entirely free of rows of minute hooks. Lateral 28 ART. 2. — H. MATSUMOTO: and ventral arm plates, as well as the skin- covered arm spines, confined to the ventral side of the arm. Vertebra) very short and exceedingly stout, discoidal, with typically streptospondyline arti- culation. Upper muscular fossas of the vertebras extremely large, the lower very small. This family includes ten genera, which may be grouped into three subfamilies as follows : Subfamily 1, Asteronychînœ mihi, 1916. — Disk large and arms slender ; only a single madreporite is present ; perihsemal canal entirely closed ; peritoneal cavity divided into five separate com- partments by the interradial attachments of the gastral pouches to the body wall ; genital bursas free from the perih^emal canal, but communicating with each other in radial pairs ; lateral arm plates of either side separated from each other by the comparatively large ventral arm plates ; more than three arm spines on each plate. Astrodia Veeeill, 1899. Aster onyx Müllek & Teoschel, 1842. Subfamily 2. TricJiasterince mihi, 1915. — Disk moderately large and arms moderately stout ; one madreporite to each inter- radius ; perihsemal canal and genital bursas with open communica- tions ; lateral arm plates of either side separated from each other by the ventral arm plates, distal ones projecting ventrally in the form of hanging rods ; two subequal arm spines on each plate. I. Arms simple. Ophiuropsis Studee, 1884. Astroceras Lyman, 1879. II. Arms divided. Trichaster Agassiz, 1835. SthenocephaÏKS Kœhlee, 1898. MONOGRAPH OF JAPANESE OPHIUROIDEA. 29 Earyale Lamakck, 1816. Subfamily 3. Aster oschematlnce milii, 1915. — Disk very small and arms very stout ; one madreporite to each interradius ; peri- hœmal canal closed ; genital bursae free from the perihsemal canal or from each other ; lateral arm plates of either side meeting each other in the ventral median line, distal ones not projecting ventrally in the form of hanging rods ; two unequal arm spines on each plate. Asteroschema Œested & Lütken, 1856.^^ Astrocharis Kœhlee, 1904. Astrogymnotes Claek, 1914. A very characteristic feature of the Asteronychiuœ is the posi- tion of the genital scales, which articulate with the genital plates near the inner ends of the latter, while in the other subfamilies, the scales articulate with the plates near the outer ends of the latter. In Aster onyx loveni Müller & Teoschel, the radial shields are composed of several lamellar secondary plates, which overlap and fuse together. The wall of the gastral cavity is attached to the body wall at the central part on the dorsal side of the disk and along the interradial lines above and below, so that the. peri- toneal cavity is perfectly divided into five radial compartments, which are filled up by the folded generative glands. The peri- heernal canal is entirely closed. The genital bursso are very spacious, but do not communicate with the perihaemal canal, and are separated fi'om the peritoneal cavity by a thin, folded membrane. Each radial pair of tlio bursa? however communicate with each other just above the outer end of the oral frames and the first 1) Including Ophiocreas Lyman, 1869. As far as the external characters are concerned, Ophiocreas adherens Studer, 1884, and Asteroschema kcehleri Dödeblein, 1898, appear to be referable to the Tr'ichasterlnœ ; they may possibly be young forms of Aslroceras, 30 ART. 2. — H. MATSUMOTO : vertebra. In very young specimens, the genital plates are very high in position, lying nearly dorso -laterally to the first two ver- tebra, and the lower muscular fossa3 of the vertebrsG are not notably smaller than the upper, so that the genital plates and vertebras rather remind one of those of the Ophiohyrsince. A most important internal character of the Trîcliasterînœ is that the genital bursas communicate with the perihasmal canal. In Astroceras, Trichaster and Euryale, I was able to observe this communication ; and I behove that this character is a decisive proof for the close relationship of these genera, as stated by Lyman and Kœhler. In Astroceras pergamena Lyman, the genital bursas communicate merely with the perihaemal canal, while in Trichaster elegans Ludwig, they communicate also with each other in radial pairs, the communications taking place above the vertebrae just inside the disk margin ; in both the genera, the peritoneal cavity, as well as the generative glands, is confined to the arm bases. The skin of the arm consists of two layers, which are easily separated by boiling in potash. The inner layei*, which is thicker than the outer, contains the rudiments of the dorsal arm plates, which are in smaller specimens very insignificant, being represented by double rows of very feeble calcareous scales, continued above fi'om the lateral arm plates, as stated by Lyman in Euryale aspera Lamaeck. But in larger specimens, with the generative glands extending into the arm bases, the dorsal arm plates become con- spicuous in the proximal arm joints containing the generative glands, so that the arm bases are strongly ribbed. At this stage, each dorsal plate is represented by a pair of bar-like plates in Astroceras, and by a pair of about triple rows of nodule-like secondary plates in Trichaster. In Trichaster elegans, the genital plates and scales are very stout, lying very closely side by side MONOGRAPH OF JAPANESE OPHIÜROIDEA. 31 to form very solid interbrachial borders. The peristomial plates are stout, without median groove, soldered with the oral frames and oral plates. A single, very long dental plate is present at the apex of each jaw. The adorai shields are very stout, divided into inner and outer halves, so that there are four in each interradius, closely set together. The radial shields are stout, composed of several secondary plates, which are soldered together. The ventral arm plates are large, often divided into two or three secondary plates. The lateral arm plates are very small, those of the two sides not meeting each other ; in the distal arm joints outside the second bifurcation, they are projected ventrally in the form of hanging rods, bearing two hook-shaped arm spines. The dorsal arm plates are represented by three or four rows of nodule -like secondary plates on either side of each proximal arm joint, two rows being internal and the other one or two external. According to Lyman, the genital shts in Euryale aspera penetrate directly into the " peritoneal cavity," which communicates with the peri- haemal canal, and the generative glands lie in the " peritoneal cavity," the " genital bursa3 " being entirely absent. As far as I can judge, Lyman seems to be mistaken in his interpretation of the parts. In my opinion, his " peritoneal cavity " is not the genuine peritoneal cavity but merely the genital bursa?, which are in direct communication with the perihsemal canal and are very spacious and provided with a very thin membrane ; and the genuine peritoneal cavity is perfectly divided into five radial compartments, which are entirely filled up by the folded generative glands. The correctness of this interpretation of mine is proved by the fact that, the generative glands do not lie in the " peritoneal cavity," as Lyman thinks, but are separated from it by a thin but distinct membrane and the brachial body cavity, which is the direct con- 32 AET. 2. — H. MATSUMOTO : tinuation of the true peritoneal cavity is entirely free from Lyman's " peritoneal cavity." Thus, it may clearly be seen that, Eiiryale is similar to Astroceras and Trichaster in having direct communica- tion between the genital burssG and the perihsemal canal, and to the Asteronychinœ in having the generative glands and the peritoneal cavity lying in the disk instead of being confined to the arm bases. The Asteroscliematince are similar to the Asteronychinœ in the perihfcmal canal being entirely free from the genital bursse, but differ from them in the genital bursas being free from each other. In Aster oschema japonicum (Kœhlee), the genital plates are long, stout, lie oblique to the arm axis ; and curve abradially in the outer half of their entire length, so as to narrow the ventral interbrachial spaces. The genital scales are very small, directed laterally to support the disk borders, instead of supporting the abradial border of the genital slits. The genital bursas are very spacious, but do not communicate with the periha3mal cavity or with each other. The radial shields are very long, bar-like, com- posed of several secondary plates, so that they are capable of bending like elastic bars. The peristomial plates are very stout, witli notched outer borders, distinctly grooved in the radial line on the dorsal surface. The oral frames are stout, more or less cylindrical, without distinct distal wings. Ludwig's ' first ventral arm plate ' is present at the dorsal outer corner of each oral slit. Oral plates strongly projecting ventrally as well as dorsally, and provided on the adradial sides with low, smooth, pavement-like grains, corresponding to the oral papilloe. The dental plates are in a single piece for each jaw, w^ith a canal between plate and jaw. Teeth very stout, arranged in a single vertical row. Adorai shields entire, very large ; the outer open angle between eacli MOXOGRArH OF JArANESI^ OrHIUROIDEA. 33 interradial pair is filled by the oral shield, the raadreporite, which is rather large and pear-shaped. The ventral arm plates are very small and grannie-like, except the first and second, which are comparatively large, the first being triangular, with the apex: directed inwards, and the second being fan-shaped with a convex, longest outer side. The lateral arm plates are transversely bar- like, meeting each other in the ventral median lino. Their outer ends are forked, the arm spines arising from the ventral process, while the row of plates which correspond to the dorsal arm plates proceed dorsally from the dorsal process. In the proximal joints of large, sexually mature specimens, each dorsal arm plate is re- presented by a pair of simple, bar-like rows of secondary plates, while in the distal joints, as throughout in smaller specimens, each dorsal plate is represented by a double row of very fine, insignifi- cant scales. In Asteroschema caudatum (Lyman), the structure is essentially similar, but the lateral arm plates are stouter than in the foregoing species, and the ventral arm plates are often divided into two secondary plates. Asteronyx lovent Müller & Troschel. Asteronyx loveni : MÜLLER & Troschel, Sys. Asfc., 1842, p. 119, PI. X, figs. 3-5; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 285; Döderlein, Abh. Math.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 115; Mortensex, Zeitsclir. wiss. Zool., CI, 1912, pp. 264-289, Pis. XIV- xvin. Astromjx loveni : Lyman, Hep. Challenger, V, 1882, p. 285 ; Kcehler, Ech. lud. Mus. Deep-sea. Opli., p. 74; Kœhler, Exp. Siboga, XLV, Pfc. 1, 1904, p. 167. One specimen ; Hokkaido, more precise locality unknown. Several specimens, clinging to Isis and other gorgonaceans ; off Misaki. 34 AKT. 2. — H. MATSUMOTO Japan ; 350 fathoms (Lyman). Hiuga Sea ; 405-578 fathoms (Clakk). Off Omai Zaki ; 475-918 fathoms (Clark). Off Smio Zaki, Sagami Sea ; 83-158 fathoms (Clark). Off Kinkwa San ; 129 fathoms (Clark). Off Kii, Kumano Sea ; 440-649 fathoms (Clark). Okhotsk Sea ; 100-510 fathoms (Clark). Indian Ocean (Kœhler). Malaysian waters (Kœhler). Bering Sea (Clark). Alaska (Clark). Wasliington (Clark). California (Clark). Eastern coasts of North America. West Indies. Nor- wegian coasts. Finmark. Scotland. The specimen from, Hokkaido is very large, and is 42 mm. in the disk diameter, 500 to 590 mm. in the arm length and 5.5 to 6.5 mm. in the arm width at the base ; the largest arm having about six hundred and twenty joints. The other specimens are all young, with the disk diameter of 6 to 7 mm., and hardly showing the specific characters of the oral ap- paratus and arni spines. The oral papillœ and teeth are flat, serrate along the free margin, covered over by skin, except at the free margin ; so that, they are quite similar to ^ A^ /^^s v^ those of Astrodia. On the disk, ^ rH^^^'^'^ . "^f,-., V there are several scattered, delicate scales embedded in the skin, which 'é ■V; ^ Fig. 6. Asieronyx loveni a. From above. xio. h. From below, xio. c. From remind US of the disk scales of below. X 11. a and h. Young. Astrodia, and are destined to be absorbed in older examples. MONOGRAPH OF JAPANESE OPHIUROIDEA. 35 Key to Japanese genera of Trichastevinœ. A — Arms simple ; generative glands extending into the &rm. bases as two pairs of ribbon-like bodies ; in basal arm joints of adult speci- mens, each dorsal arm plate is represented by a pair of transversely bar-like plates, which bear a stumpy tubercle at the upper end . . Asiroceras. AA — Arms dichotomously divided. a — Arms divided only near the tip ; generative glands extending into the arm bases as four pairs of ribbon-like bodies ; in basal arm joints of adult specimens, each dorsal arm plate is represented by a pair of about triple rows of nodule-like plates TricJiaster. aa — Ai*ms divided nearly from the base ; generative glands confined to the disk ; dorsal arm j)l^tes always very insignificant Euryale. Astroceras ijergainena Lyman. Astroceras percjamena : Lyman, Bull. Mus. Comp. Zool., VI, 1879, p. 62, PL XVin, figs. 478-480; Lyman, Kep. Challenger, V, 1882, p. 284, PL XXXIV, figs. 1-5; Clark, Zool. Anz., XXV, 1902, p. 671; Kœhler, Exp. Siboga, XLV, Pt. 1, 1904, p. 159; Clark, BulL U. S. Nat. Mus., LXXV, 1911, p. 284; Döderlein, Abh. Math.-Phys. KL K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 61, PL VI, figs. 4-4 &, PL Vm, fig. 13. AsiroscJiema sp. ; Döderlein, ibid., p. 57, PL VII, fig. 3. Numerous specimens ; Sagami Sea ; 100 fathoms. Numerous specimens ; off Ukislaima, Uraga Channel ; 300 fathoms. Eastern Sea ; 95-106 fathoms (Clakk). Yenshii Sea ; 565 fathoms (Lyman). Off Omai Zaki ; 34-37 fathoms (Clark). Suruga Gulf ; 52-180 fathoms (Claek, Döderlein). Timor ; 216 m. (Kœhler). The largest one of my specimens is 7 mm. across the disk. The first ten basal free joints of the arms are wide and provided with transverse plates, which form ridges on either side and are 36 AET. 2. — H. MATSUMOTO well developed in every second or third joint, in which they bear each a conspicuous, movably articulated tubercle at the upper end. Most specimens are quite young and srnaller than 4 mm. in the disk diameter. They are mostly six -armed, rarely five or seven-armed. On the upper side of each arm joint, there are two transverse rows of rough, skin-covered nodules, six to eight in each row. Similar nodules occur on the radial ribs ; and in the larger specimens they are also found on the joints of the slender parts of the arms. I have also some six- armed specimens, in which the middle one of the three regenerating arms is distinctly smaller than the other two ; but I do not agree with Clark in looking upon the smaller arm as being destined for resorption, so that the animal would become five-armed. The smaller size of the arm in question is perhaps due to its unfavourable position with regard to the supply of nutrition. In fact, we see in many ophiurans which reproduce by schizogony, that the regenerating arms do not often keep to the original number, and tlie same occurs also in Fig. 7. Astroceras ■pergamemt. a. From above. X 7. h. From above, x 8. c. From be ow. x 14. h and c. Young. MONOGRAPH OF JAPANESE OPHIUROIDEA. 37 this species. I have -five-armed specimens, in which two of the arms are smaller and are undoubtedly regenerated ones ; so that, the five-armed condition may be attained by the final regeneration of only two arms in place of three, in the equally divided halves of six-armed individuals. Key to species of Trichaster. A — Interbracliial ventral surfaces extremely narrow, so that the two genital slits are so close together as to form a single ajperture ; arms almost triangular in section ; about sixty-two arm joints within the first bifurcation ; each proximal arm joint bearing a pan' of stumpy tubercles on the dorsal side ; arm width at the first bifurca- tion being about one-third of that at the base palmiferus. AA — Interbrachial ventral surfaces not extremely narrow, so that the two genital slits are rather well separated from each other ; arms quad- rangular in section ; thirty-nine to fifty-three, usually forty-five, arm joints within the first bifurcation ; arms entirely free of stumj)y tubercles; arm width at the first bifurcation being less than one- fourth of 'that at the bass elcgans. Trichaster pahnifertis (Lamarck). Euryale palmiferum : Lamarck, Hist. Nat. Anim. sans Vert., II, 1816, p. 539. Trichaster palmiferus : Agassiz, Mem. Soc. Sei. Nat. Neuchatel, I, 1835, p. 139^^ ; MÜLLER & Troschel, Sys. Ast., 1842, p. 120 ; Lyman, Rep. ChaUenger, V, 1882, p. 267; Clark, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 287 ; DÖDERLEIN, Abh. Math.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.- Bd. I, 1911, p. 62 ; Bomford, Bee. Ind. Mus., IX, Pt. 4, 1913, p. 220, PI. xni, fig. 2. Trichaster frageUifer : V. Martens, Wieg. Arch. Nat., XXXH, 1877, p. 87.^^ 1) These papers were not seen by me. 38 ART. 2. — H. MATSUMOTO ! Colnett Strait ; 83-84 fathoms (Clark). Bay of Bengal. Ceylon. Malaysian waters. Hong-kong. Trichaster elegans Ludwig. Tricliaster elegans : Ludwig, Zeitschr. wiss. Zool., XXXI, 1878, j^p. 59-67, PL Y, figs. 1-9 ; Lyman, Eep. Challenger, V, 1882, p. 267 ; Bomford, Rec. Ind. Mus., IX, Pfc. 4, 1913, p. 220, PI. XHI, figs. 3 & 4. Two specimens ; Tanabé Bay, Kii. Pacific Ocean (Ludwig). India (Bomfoed). DöDERLEiN thinks that the present species is a young form of T. jKilmiferus ; but like Bomford I can not agree with this opinion. My two specimens are quite large, and are of course adult and sexually mature. The large arm tubercles, which are characteristic of T. Fig. 8. TriclwLster elegans. xl. a. From above, h. Ventral view of skeleton of two arm joints, c. Frontal view of skeleton of an arm joint, d. Lateral view of skeleton of two arm joints. EX. External row of nodules, ind. Internal distal row of nodules, inp. Internal proximal row of nodules, lp. Lateral arm plate, vp. Ventral arm plate. MONOGRAPH OF JAPANESE OPHIUROIDEA. 39 palmiferus, at least in the adult, are entirely absent in my specimens. DoDERLEiN states that, the part of the arm within the first bifurca- tion is composed of about sixty-two joints in T. palmiferus, wliile in my specimens, that part is composed of from forty-three to forty-five joints, except in an arm of one of the specimens, in which that part is composed of fifty-three joints. Ludwig makes no state- ment on this point, but in his fig. 1, about forty-six joints are indicated in the part in question. This number almost agrees with what obtains in my specimens, but is very far from that given by Düderlein. Bomford records thirty -nine to fifty -two joints in the part in question. So I must agree with Bomford in recognising the present species as distinct from 2\ palmiferus. In one of my two specimens, the disk is 33 mm. in diameter, and the arms are divided six times and about 310 mm. in total length. In the other specim,en, the disk is 32 mm. in diameter, and the arms are divided five or six times and about 290 mm. in total length. As to the ratio of the width to height of the arms the two specimens stand in striking contrast, as shown in the followincc table. Disk tliameter of specimen. Width and height of arm at base. Ditto at the 20th. free joint. Ditto at the 40th. free joint. 33 mm. W. 17 mm. H. 15 mm. W. 12 mm. H. 13 mm. W. 7.5 mm. H. 8 mm. 32 mm. W. 17 mm. H. 13 mm. W. 13 mm. H. 12 mm. W. 8 mm. H. 7.5 mm. The generative glands extend into the arms as four pairs of ribbon-shaped bodies for each arm. Microscopical examinations of the generative glands showed that, the larger one of the two specimens was male and tlie smaller one female. The testicular 40 ART. 2. — H. MATSUMOTO : « ribbons are much more slender than the ovarial ; so that, the dorsal median groove of the arms is much deeper and narrower in the male than in the female. I am inchned to think that, the striking contrast of the ratio of the width to height of the arms in the two specimens is a sexual difference. I have one more statement to make about the smaller speci- men, viz. that it is four-armed. Not only the arms but the body generally is quadrimerous throughout, and there are no indications of external injury. The great width of the arms in this specimen might be looked upon as a result of the quadrimerism ; but I think it only explains the great width of the arm bases, and not that of the distal part of the arms within the first bifurcation. Key to species of Muryale. A — Proximal parts of arms beset with two rows of large stumpy tuber- cles ; some similar tubercles are found also on the radial ribs aspera. AA — Arms entirely free of stumpy tubercles, but some very small tuber- cles j)i'esent on the radial ribs anopJa. Euryale aspera Lamaeck. Euryale aspevum : Lamarck, Hist. Nat. Anim. sans Vert., II, 181G, p. 535. Aslropliyton aspcriim : Müller & Troschel, Sys. Ast.. 1842, p. 124 ; Ludwig, Zeifschr. wiss. Zool., XXXI, 1878, p. 66. Enrijalc asp>era : Lyman, Piop. Challenger, V, 1882, p. 266 ; Studer, Abh. K. Akad. Wiss. Berlin, 1882, p. 53, PI. Y, fig. 10 ; Kœhlee, Bull. Sei. Fr. Belg., XXXI, 1898, p. 114. Euryale studer! : LoRiOL, Piev. Suisse Zool., VIII, 1900, p. 8, PI. VIII, fig. 4, PI. IX, fig. 1 ; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 132 ; Kœhlee, Bull. Sei. Fr. Belg., XLI, 1907, p. 350. MONOGRAPH OP JAPANESE OPHIUROIDEA. 41 Euryala aspera : Döderlein, Abh. Matli.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 65 & 98, fig. 16, PL V, figs. 7 & la. One specimen ; Okinawa. Eastern and western Australia. Malaysian waters. Philippines. Gulf of Siam. Southern China. This specimen is very small, being only 4 mm. in the disk diameter. The arms are branched about eight times, and are about 27 mm. in total length. At the cen- tral region of the disk, six very small, rather inconspicuous primary plates are present. Each radial rib bears near its outer end one large, cylindrical, blunt tubercle joined by a movable articulation. A grain is present in the radial line somewhere between each pair of the tubercles. Further, some of the arm joints bear each one short, cylindrical, blunt tubercle joined by a movable articulation. Such a tubercle never lies in the median line, but in such a position as to suggest that it is one of a pair, of which the fellow is absent. Fig. 9. Euryale aspera. a. From above. x2. h. From belo'«'. x 10. JEurydle anopla Clark. EunjaU anopla: Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 294, fig. 144. 42 ART. 2, — H. MATSUMOTO : Euryala anopla : Döderlein, Abb. Math.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 98. Eastern Sea ; 103-152 fathoms (Clakk). Key to Japanese genera of Asteroschematinœ. A — Eadial shields entirely covered over ; interbrachial ventral surfaces not so narrowed as to be slit-like ; arm bases strongly ribbed in adult specimens ; arm spines not very small Aster oscliema. A A — Kadial shields partially naked ; interbrachial ventral surfaces very narrow and slit-like ; arm bases much widened, not strongly ribbed, but almost smooth ; arm spines exceedingly small Astrocharis. Key to Japanese species of Asteroscliema, A — Disk and arms closely covered with granules . . typical Asteroscliema. a — Kadial ribs narrow, not occupying the entire dorsal side of disk ; arms about fourteen times as long as the disk diameter ; three or four granules in 1 mm. on the dorsal side of the arm bases ; oral tentacle pores and some ten &st tentacle pores opening by means of tubes tuhi/erum. aa — Kadial ribs stout, almost occupying the entire dorsal side of the disk ; five or six granules in 1 mm. on the dorsal side of the arm bases ; oral tentacle pores and only a few first tentacle pores open- ing by means of tubes. h — Ventral side of arms closely covered with granules, like the other parts ; arms about nine times as long as the disk diameter glaucum. hb — Ventral side of the arms covered by a skin, which is almost free of granules ; arms about twelve times as long as the disk diameter hemigymnum. AA — Disk and arms covered by a thick, naked skin, which may, how- MONOGRAPH OF JAPANESE OPHIÜEOIDEA. 43 ever, contain fine granules in younger stages .... section Ophiocreas. c — Arms one-tliird to one-fourth as wide as the disk diameter. d — Arms thirteen to seventeen times as long as the disk diameter, e — Skin of the disk and arms very thick ; arm spines not very slender ; no large depressions just outside the tentacle pores caudatum. ee — Skin of the disk and arms very thin ; arm spines long and very slender; a large depression present just outside each tentacle pore. , japonicum. dd — Arms about eight times as long as the disk diameter; skin of the disk and arms moderately thick ; arm spines short ; genital slits very short, situated near the interbrachial angle ahyssicola. cc — Arms about one-half as wide as the disk diameter, and nine to twelve times as long as the same ; skin of the disk and arms thick, covering the underlying parts very loosely ghdinosum. It is by no means easy to distinguish clearly the species of Asteroschema, especially those of the section Ophiocreas, owing to the fact that, this genus assumes different appearances according to its developmental stages. The oral papillae are papilliform and arranged in a single horizontal row in younger stages, but are con- verted into flattened and pavement-like grains in larger specimens. The generative glands extend into the arm bases as two pairs of ribbon- shaped bodies only in the adult, the extent of the lobes increasing with gi'owth. The parts of the arms containing the generative glands are much widened and knotted, owing to the appearance of a transverse row of secondary plates on either side of each arm joint. In the section Ophiocreas, the younger speci- mens are more granular, and the full-grown ones are almost entirely naked. The ratio of the arm length to the disk diameter increases with growth. In exceedingly young specimens of a cer- tain species, only a single arm spine (the adradial one in the larger specimens) is present for each tentacle pore. On the basis u ART. 2. — H. MATSUMOTO of these considerations, I am convinced that, the identification and naming of Aster oschema merely on the ground of the oral papillae, widened or not widened arm bases, degree of granulation, arm length, &c., without reference to the size of the specimens, would inevitably lead to a great confusion. Thus, I am obliged to transfer several species set up by eminent authors to the list of synonyms. Asteroschcnia tiibiferutu MatsUxMoto. Asteroschema tuhifermn : Matsumoto, Proc. Acad. Nat. Sei. Philadelpliia, 1915, p. 52. Two specimens ; Okino-sé, Sagami Sea. One specimen ; off Misaki, Sagami Sea. Diameter of disk 16 mm. Length of arras 230 mm. Width of arms at base 4.5 mm. Disk rather arched, a little higher than the arms, closely covei'cd with very fine smooth granules. Radial ribs long and nar- row, narrower within, nearly reaching the centre. The disk granules are finer and rather well spaced towards the centre, but coarser and very close- set on the radial ribs. Interbrachial ventral sur- faces rather vertical, form- Kig. lO. Asteroschema iuhiferum. x4. a. From above. &. From below. ing a notch, in which lie MONOGEArH OF JAPANESE OPHIUEOIDEA. 45 tlio two rather large, nearly parallel genital slits. On the floor of each notch, opens one madreporic pore. Oral angles convex late- rally and downwards, almost filhng up the oral slits. Eleven or twelve teeth in a single vertical row, triangular and stout. On either side of the oral angle, there are several flat, smooth, pave- ment-like grains corresponding to the oral papillae. Arms long and slender, as high as wide near the base, not smoothly rounded, but inclined to be quadrangular in transverse section, wider above than below ; more distally arched above and plane below, higher than wide ; closely covered with fine, smooth granules, of which three or four lie in 1 mm. on the dorsal side near the arm base. Arm joints rather distinct. First tentacle pore without arm spines, next three or four pores with one, and the rest with two spines. The abradial spine is very small, conical, enclosed in skin, with rough end. The adradial one is short and conical in the basal joints ; more distally they become longer, stouter and club-shaped, finally about as long as two arm joints ; enclosed in skin, rough ; the minute thorns at the end of the spines become distally rather concentrated on the inner side of the spines. Oral tentacle pores and some ten basal tentacle pores opening by means of tubes, in which they are enclosed ; each tube, except that of the oral and the first tentacle, is attached to the adradial arm spine on its adradial side. Colour in alcohol light pinkish brown or flesh-coloured. This species is quite near to A. rubrum Lyman, but differs from it cliiefly in the much coarser granules on the basal parts of the arms, in having tentacle tubes not only for the oral tentacles but also for some ten first tentacles, and in the club-shaped and rela- tively longer and stouter arm spines. 46 ART. 2. — H. MATSUMOTO Astevoscheina glaiicuin Matsumoto. Asteroschema glaucum : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 53. Three specimens; Dôketsuba, Sagami Sea; 110 fathoms. Diameter of disk 11mm. Length of arms 100 mm. Width of arms at base 4 mm. Disk five-lobed, the lobes being continued without any distinct de- marcation into the arms ; flat, about as liigh as the arms, closely cover- ed with very fine, smooth granules. Kadial ribs rather indistinct. In- terbrachial ventral surfaces rather vertical, very narrow, forming a deep notch, in which lie the two rather large, parallel genital slits. Oral angles convex laterally and downwards, almost filling up the oral slits. Six or seven teeth in a single vertical row, triangular and stout. On either side of the oral angle, there are seve- ral flat, smooth, pavement-like grains corresponding to the oral papilla?. Sometimes a few round smooth grains occur below the teeth. Arms as high as wide, very stout at the base, so that the interbrachial spaces ai^e very narrow below ; distally rather slender and higher than wide ; covered with fine, smooth, close-set granules, of which there are about six in 1 mm. on the radial ribs and the arm bases. Arm joints in- distinct in the basal parts. First tentacle pore w^ithout arm Fig. 11. Asteroschema glaucum. x4. «. From above. h. From below. MONOGKAPH OP JAPANESE OPHIUROIDEA. 47 spines ; next four or five with one, and the rest with two spines. The abraclial spine is very small, conical, enclosed in skin, with rough end. The adradial one is small and conical in the basal joints, distally longer and somewhat clnb-shaped, and slightly longer than the. arm joint itself ; enclosed in skin, rough ; the minute thorns at the end of the spine become distally rather concentrated on the inner side of the spine, and finally, towards the very extremity of the arms, the arm spines are transformed into compound hooks, each with three or four booklets. Oral tentacles enclosed in tubes. The first two or three tentacle pores are also provided with tubes, though rudimentary. Colour in alcohol pale gray. This species is very near to A. salyx Lyman, but difiers from it chiefly in the coarser granules of the disk and arm bases, in the stout arm bases being as wide as high, in the much shorter arm spines, and in the oral tentacles being enclosed in tubes. Astevoscherna hemigyinnuin Matsumoto. Asteroschema lieimgymnum : Matsumoto, Proc. Acad. Nat. Sei. Phila- delphia, 1915, p. 53. One specimen ; off" Misaki, Sagami Sea. Diameter of disk 10 mm. Length of arms 120 nun. Width of arms at base 3 mm. Disk rather flat, divided into ten lobes corresponding to the radial ribs by ten radiating furrows, covered by a skin, which contains very fine, smooth, close-set granules. Interbracliial ven- tral surfaces rather vertical, narrow, forming a deep notch, on the floor of which opens one madreporic pore. Genital slits rather short, a little diverging dorsally. The ventral surface of the disk is covered by a finely and rather sparsely granulated skin. Oral 48 ART. 2. — H. MATSUMOTO W-'É^^^^:"^ angles not markedly set off from the outer parts. Six or seven teeth arranged in a. single vertical row, very stout. On either side of the oral angles, there occm- several coarse, flat, smooth, pavement-liko grains corresponding to the oral papillœ. Arms very stout for the first three or four free joints, but becoming rather slender further out ; their width just outside the fourth free joint is 2.5 mm. They constantly taper outwards, so that they are exceedingly slender to- wards the extremities, and acute at the tips. Dorsal and lateral surface of the arms covered by a skin, which is similar to that of the disk and contains very fine, smooth, close-set granules, of which there are about five in 1 mm. on the dorsal surface of the arm bases. The granules become much finer outwards, and almost disappear nearer the extremity of the arm. The vertebra} are visible through the skin, but the surface of the arms is practically smooth and without distinct demarcations of the joints, except in the first three or four free joints, which are marked off by shallow constrictions. The ventral surface of the arms is entirely naked, and the lateral and ventral arm plates are clearly visible through the skin. First tentacle ]Jore without arm spines ; next four or five with one, and the rest with two spines. The abradial spine 'CM Fig. 12. Asteroschema hemigymnmn. xG. a. From above, h. From below. MONOGRAPH OF JAPANESE OPfflUKOIDEA. 49 is very small, cylindrical, enclosed in skin, more or less rough at the free end. The adradial one is club-shaped, enclosed in skin, very rough at the free end. The arm spines are largest at the middle of the arms, the adradial one being one and a half times as long as, and the abradial one a little shorter than, the cor- responding arm joint. They are transformed into compound hooks with three to six booklets towards the very extremity of the arm. The oral tentacle pore and the first three or four tentacle pores are provided with tubes. Colour in alcohol grayish brown. Like A. intectum Lyman and A. migrator Kœhlee, this species appears to be an intermediate form between the sections Asteroschema s. str. and Oplnocreas. Asterosche/ina (Ophiocreas) caiidaficni (Lyman). OpUocreas caudatus : Lyman, Bull. Mus. Comp. Zool., VI, 1879, p. 64, PI. XVI, figs. 439-442; Lym.\n, Eep. Challenger, V, 1882, p. 281, PI. XXXn, figs. 5-8. Ophiocreas œdipus : Claek Bull. U. S. Nat. Mus., LXXV, 1911, p. 283. (Non Lyman, 1879.) Asteroschema {Ophiocreas) sagamimim : Döderlein, Abh. Math.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 60, PI. VI, figs. 6 & 6 a, PI. Vn, fig. 10. Asteroschema {Ophiocreas) caudcdum : Döderlein, ibid., p. 113. One specimen ; Otaba, Sagami Sea ; 420 fathoms. Numerous specimens ; outside Okinosé, Sagami Sea ; 330 fathoms. Numerous specimens ; Sagami Sea. These specimens range from very small to very large ones. In the largest one, the disk is 25 mm. in diameter, and the arms are 420 mm. in length and 7.5 mm. in width. The oral tentacle 50 ART. 2. — H. MATSUMOTO 'kP /# i3: #^ . 13ores and somo ten basal tentacle pores open by means of tubes, a character which was over- looked by Lyman. Colour in alcohol, light pinkish brown, or dark brownish purple, if the specimens have not lain in alcohol for a long time. The dermal granules, which are present in the arm bases and the disk, are more con- spicuous in smaller specimens. I have failed to find any impor- tant distinguishing characters between A. sagaminum Döder- LEiN and the young specimens of corresponding size of the present species, so that I consider tlie former as a synonym of A. caudatum. In this species, the ribbon-shaped generative glands reach out to an enormous extent. In a specimen with the disk diameter of 25 mm., they extend outwards about 180 mm. and as far as the seventieth free arm joint. The parts of the arms containing the generative glands are quadrangular in transverse section, covered dorsally by a much wrinkled skin, and have, on the dorsal and lateral sides of each joint, a row of plates supporting and protect- ing the cavity in which the generative glands lie. These plates are replaced by double rows of minute scales in tlie sterile parts of the arms. In smaller specimens, the generative lobes are con- fined to the short basal portion of the arms, winch is much widened. Thus, specimens of about 12 mm. in the disk diameter ■ Pig. 13. Asteroschenid {Oj;/( to crews) cuudatum x4. a. From above, h. From below. :\r()\oaRArir of .iapanese ophiuroidea. 51 much rosomble the typo of A. œdljms (Ly:\[an), but aro distiiignisli- od froui it chiefly by tlic presence of abundant dermal granules, by the nm(îh shorter and stouter arms, and by the tentacle pores with a single spine extending further into the arms. Through Dr. IT. L. Oi.aiîk's kindness, I was able to examine a specimen of his " Ophiocreas œdlpus'' from Japan. It is, in my opinion, undoubtedly i-eferable to the present species, so I have no hesitation to drop /Uteroschema [Ophiocreas) œcllpus from the list of Japanese ophiurans. Asteroschema (Ophiocreas) Japoniciini (KrEHT.ER). Ophiocreas japonicus : Kœhlkk, Bull. Sei. Fr. l>elg., XT J, 1907, p. .346, PL XIV, fig. 54. OjMocrcas papiUafus : Ct>ai!K, lînll. Mns. Comp. Zool., U, 190S, p. 298. ^isteroscliema {Opldo<'reas) jctponlcus : Düderlein, Abh. Muth.-riij-s. Kl. K. Bayer. Akad. AViss., Snppl.-Bd. I, 1911, p. 59, Bl. VI, figs. 5 k ^a, PL VII, fig. 9. Asteroscliema (Ophiocreas) monacanthum : ibid., p. 58, PI. AT, figs. 9-9 &. Asierosckema (^Ophiocreas) enoshimanum : ibid., p. GO, PL AT, figs. S & 8r/. One specimen ; Suruga Gulf. Three specimens ; Okinosc, Saga- mi Sea ; 200 fathoms. Numerous specimens ; outside Okinosé ; 330 fathoms, Numerous specimens ; Sagami Sea. These specimens range from very small t(,) very large ones representing all growth stages. In the larger ones, the skin is entirely free of granules. In somewhat smaller specimens, the skin contains some granules in the disk and arm bases, while in still smaller ones, the skin is strongly granulated on the dorsal, as well as on the ventral, surface of the disk and arms. The oral 52 AET. 2. — H. ÄIATSUMOTO paj)illa3 appear to have been overlooked by Kœhlek and Dödekleix, but they are present in the form of several flat, smootli, pave- ment-hke grains, which are however not veiy distinct in larger specimens, bnt are more prominent and papilliform in smaller ones. Kœhlee, Clakk and Dödeelein er- roneously call the depression just outside the tentacle pore large tentacle pore, but this large depression is distinct from the pore itself. Dödee- lein doubts the taxonomic value of this depression, but it is present in all my speci- mens and in no other species in my knowledge. I have examined pieces of arms of the foregoing and the present species after boiling them in potash. In A. caudatum, the bar-like lateral arm plates are stout, with much widened adradial ends, and the ventral arm plates are divided into right and left halves, so that the spaces for tlie tentacles are very small. In A. japon'icum, the bar-like lateral arm plates are very slender, with feebly widened adradial ends, and the ventral arm plates are very small and granuliform, so that the spaces -for the tentacles are very large. Thus, I have no doubt that, the depression in question of the present species is a specific character. I have compared Ophiocreas i^cipillatus Clakk, A. enoshimanum Dödeelein and A. monacanthum Dödeelein with the corresponding Picf. 14. AsteroscJienm {Ojjhiocreas) jajjonlmm. xi a. From above, h. From below. :\tONOGiîAPH OF JAPANESE OPHIUROIDEA. 53 specimens of the present species, and I am quite satisfied as to their specific identity. The three forms just mentioned are, in my opinion, merely younger stages of the present species before the disappearance of the dermal granules and the occlusion of the oral papillœ ; A. monacanthum, being the youngest, is only 4 mm. across the disk and has only the adradial arm spine. In slightly larger specimens, the abradial spine begins to appear as a very small rudiment. The specimens are light pinkish brown or flesh-coloured in alcohol, or dark brownish purple, if they have not lain in alcohol for a long time, or sometimes yellowish brown, when badly pre- served in weak alcohol. Asteroschenia (Ophiocreas) abyssicola (Lyman). Opldocreas ahyssicola : Lyman, Bull. Mus. Comp. ZooL, "VI, Ft. 2, 1879, p. 64, PI. XVn, figs. 470-473; Lyman, Eep. Challenger, V, 1882, p. 282, PI. XXXn, figs. 1-4. Asteroschema {Ophiocreas) ahyssicola: Dodeklein, Abh. Math.-Pliys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 113. Eastwards from Honshu ; 2,300 fathoms (Lyman). Asteroschenia (Ophioci^eas) glutinosum Doderlein. Asteroschema ( Ophiocreas) (jlutinosimi : Doderlein, Abh. Math.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd., I, 1911, p. 59, PI. VI, figs. 5 & 5 a, PI. VII, fig. 9. One specimen (belonging to the First High School) ; Sagami Sea. Numerous specimens, clinging to a gorgonacean, together with some specimens of A. caudatum and A. japonicum ; outside Okinosé, Sagami Sea ; 330 fathoms. The ratio of the arm length to the disk diameter is much 54 ART, 2. — H. MATSUMOTO smaller in smaller speci- mens. Three of the above specimens are six-armed, but there is no trace of schizogony. On either side of the oral angles, there are several flat, smooth pavement-like granules, 'which are however not very distinct. Sometimes, there occur one or two papilliform grains below the teeth. The colour is white in alcohol, or brown, if the specimens have not lain in alcohol for a Ions time. Fig. IS. Asteros-chemn (Ophiocreds) glutlncmmi. a. ^^^ AstrOCei'US, AsteiO- From above. x§. h. From nbove. x|. c. From schema and TricIlClster, the bf'lo^\'. X 3. generative glands extend into the arms as ribbon- shaped bodies, and as a consequence, the arms present certain peculiar features. The parts of the arms containing the generative lobes are more or less widened, quadran- gular in transverse section," strongly knotted, and has a conspicu- ous dorsal median groo^e. In Astroceras and Aster oschema, there arc two pairs of generative glands in each arm, and tlie parts containing them present a knotted appearance, owing to the pre- sence of transverse ridges on either side. In .the first named genus, these ridges are formed by bar -like plates, but in the latter MONOGRArn OF JAPANESE OPHIUROIDEA. 55 by a row of plates, supporting and protecting the cavity «in which the generative glands lie. In the sterile parts of the arms, these plates are replaced by two rows of nodules in each joint in Astroceras, and by two rows of minute scales in Asteroschema. In Trichaster, there are four pairs of generative glands in each arm, which bears masses of irregular nodules. These nodules are ar- ranged in two la}T'rs and form three or four rows on either side of each joint in the proximal parts of the arms : two of the rows belonging to the internal layer, two to the external layer on the ventral side and one on the lateral side. Besides, there is a large nodule on the upper end of the ridge formed by the rows of nodules on either side of each joint. These nodules form two longitudinal rows of humps on each arm. In the sterile part of the arms, the nodules are replaced by two rows of scales in each joint. The species in which the extension of the generative glands into the arms has been observed are shown below, together with some additional facts. Species. Disk diameter of specimen. Free basal arm joints contmning the generative glands. Length of the foregoing part. Astroceras pergamena Asteroschema tuhiferum 7 mm. 16 mm. 10 8 9 mm. 15 mm. „ glaiicnm 11mm. 0 0 „ hemigymnum „ caudatwn 10 mm. 25 mm. 4 70 5 mm. 180 mm. '» j> 12 mm. 6 10 mm. „ japomcum 31 mm. 46 120 mm. „ glutinosiün 22 mm. 25 mm. 10 12 20 mm. 25 mm. Trichaster elegans 15 mm. 33 mm. 32 mm. 36 15 mm. 165 mm. 120 mm. 56 ART. 2. — H. MATSUMOTO Astrocharis ijimai Matsumoto. Astrocliaris ijimai : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia 1915, p. 54. Numerous specimens, clinging to a coral ; Misaki, Sagami Sea. Diameter of disk 4.5 mm. Length of arms 50 mm. Width of arms at base 2.5 mm. Disk five-lobed, with deeply indented interbrachial borders, with the lobes emarginate towards the arms, flat, sunken at the central region, raised at the lobes, covered with very fine, smooth, irregular scales, which are very closely set and partly imbricated. Fig. 16. Astrocharis ijimai. a. From above. x2. b. From below. xlO. Eadial shields naked, very small, triangular, with the apex turned within, tuberculous when examined under the microscope. Inter- braehial ventral surfaces forming very deep notches, exceedingly narrowed by the very wide arm bases. Two genital slits, small, parallel, nearly vertical. On either side of each lobe of the disk, MONOGRAPH OP JAPANESE OPHIUROIDEA. 57 lies the naked genital plate, which is large, oval, and tuberculous under the microscope like the radial shields. Thus, each inter- bracliial ventral surface is bounded on either side by the genital plate. Oral angles puffed laterally, almost fiUing up the' oral shts. Teeth small, triangular, arranged in a single vertical row. No oral or dental papilla?. Arms very wide at the base, keeping the same width to the distance of about 4 or 5 mm., then rather rapidly narrowed, be- coming slender and cylindrical, with the width of about 1 mm. ; covered with scales similar to those of the disk. Arm joiuts in- visible in the proximal part, but more or less distinct distally. First tentacle pore without arm spine, the following ones with one spine, which is very small, short, peg-hke, somewhat flattened, rough at the end under the microscope, lying flat on the ventral surface of the arm. Half way out in the arms, the tentacle pores are each provided with two spines, of which the second or the abradial one is exceedingly small and rather inconspicuous ; while the adradial one is a little longer and erect. Colour in alcohol white or light pale yellow. In smaller specimens, the arms are scarcely widened at the base, which is also the case in regenerating ones ; for, schizogony takes place in this species, as in the genotype, A. virgo KœiTlek. Most specimens are five-armed, but the arms are often unequal, two or three being larger than the others. I have, however, one specimen with six arms, three larger and three smaUer. In the four-armed specimens, two or three of the arms may be larger and the other two or one smaller. Still another specimen has only three arms and three pairs of radial shields, doubtless in- dicating that it has lately undergone division, and that the lost parts have not been regenerated. It can be easily observed in 58 . ART. 2. — H. MATSUMOTO r this species that, only two arms may sometimes be regenerated in place of three, or one in place of two. This species can be distinguished very easily from the geno- type, A. virgo. The arms are slightly less widened at the base and keep their width for a less extent. The arm spines are only one to each tentacle pore for over half the length of the arms, while in the genotype, two are present even in the basal region. Finally, the disk and arms are covered with fine scales instead of granules. In this species, the widening of the arm bases is caused by the widening of the vertebra?, and is not accompanied b}^ the ex- tension of the generative glands, so that even in the widened part, the arm covering of the dorsal side is in direct contact with the vertebrae. I suppose that this condition is a generic character, though there is no observation on this point in the genotype. Family 3. Gorgonocephalidse (Dödeeleix, 1911) mihi, 1915. Disk and arms covered by a thick skin, which contains granules, very often beset with stumpy tubercles. Eadial shields long, bar-like, forming a system of radiate ribs. The radial shields and genital plates articulate with each other by means of a trans- verse ridge on both the plates. Genital plates long and stout, genital- scales small. Oral and adorai shields very small, the former being often separated from the latter by a mosaic of sup- plementary plates. Teeth and dental papilla?, often as well as oral papilla?, similar, spiniform, acute, forming a clump or cluster at the apex of each jaw. Peristomial plates entire, stout, soldered w^ith the also very stout oral frames. Arms exceedingly long, simple or branched, vertically coiled. Dorsal arm plates represent- MONOGEArH OF JAPANESE OrHIUROIDlvV. " 59 ed by double rows of liook-bearing granules, so that the arm is annulated by the minute hooks. Lateral and ventral arm plates confined to the ventral side of the arm. More than three arm spines, all ventral in position, serving as tentacle scales. Vertebrae very short and exceedingly stout, discoidal, with typically strepto- spondyhne articulation. Upper muscular fossos of the vertebrae ex- tremely large, the lower very small. This family includes twenty- four genera, which may be grouped into two subfamilies. Subfamily 1. Gorgonocephalince mihi, 1915 : — Teeth, dental papillae and oral papill£e all similar, spiniform ; oral angles not strongly projected ventrally ; genital slits small, often pore-like, lying near the disk border ; basal vertebrae not very small, not covered over by the muscles lying between the basal vertebra? and genital plates ; arms simple or divided ; those forms with arms simple or divided a few times have no supplementary plates in the spaces between the oral angles and the interbrachial ventral surfaces, while those with arms divided many times have well developed supplementary plates in the spaces mentioned. I. Arms simple or divided a few times ; no supplementary plates in the spaces between the oral angles and the interbrachial ventral surfaces. a. Arms simple. Astrogomphiis Lyman, 1869. ' Astrochlamys Kœhlek, 1912. Astrochele Veekill,- 1878. Asteroporpa Œested & Lütken, 1856. h. Arms divided a few times. Astrocnida Ly]\ian, 1872. Conocladus Clx\ek, 1909. II. Arms divided many times ; supplementary plates well 60 ART. 2. — H. MATSUMOTO : developed between the oral angles and the interbrachial ventral surfaces. a. Madreporite single. Astroconus Dodeelein, 1911. Gorgonoceplialus Leach, 1815. Astrodendrum Dodeelein, 1911. Astrocladus Veeeill, 1899. Astrospartus Dodeelein, 1911. Astroboa Dodeelein, 1911. Astrophyton Müllee . Arms divided a few times. Astrocolon Lyman, 1879. Lyman, ^yl^oso erroneous interpretation of certain parts in Euryale aspsra was pointed ont before, has made a similar mistake in the present family, inasmuch as he states that, in Astrocnida isiclis (Duchassaing), Gorgonocephalus arcticus Leach \_:=Ct. agasski (Lyman)] and G. chilensis (Philippi), the genital slits are not sur- rounded by any " bursa " but directly penetrate into the " peri- toneal cavity," in which the generative glands lie. I have studied the internal structure of Astrochele lymani Verrill, Asteroporpa hadracantha Clark, Gorgonocephalus tuberosus Döderlein, G. arcticus, G. caryl (Lyt^ian), G. dolichodactglus Döderlein, Astrodendrum saga- minum (Döderlein), Astrocladiis annulatus Matsumoto, A. coniferus (Döderlein), Astrohoa arctos Matsumoto, Astrothamniis echinaccus Matsumoto and Astrotoma sohrina Matsumoto, and am convinced that, Lyman's " peritoneal cavity " is not the genuine peritoneal cavity but merely the genital bursa), which are very spacious and communicate with each other, and usually also with the perihœmal canal, so as to form a single compound cavity ; the genuine peri- toneal cavity is perfectly divided into five radial compartments, which are entirely filled up by the very voluminous, folded gene- rative glands, and no longer appear as cavities ; the generative glands do not lie in Lyman's " peritoneal cavity," i.e. genital bursa?, but are separated from them by a thin but distinct membrane and lie morphologically outside them. As I have already pointed out in Euryale aspera, the fact that the brachial body cavity, which is the direct continuation of the true peritoneal cavity, has no direct communication with Lyman's " peritoneal cavity," 62 ART. 2. — H. MATSUMOTO: i.e. genital burscx\ also in these genera, is a decisive proof against Lyman's interpretation. According to Lyman, AstrogompJius vaUatiis Lyman has a very shnple oral skeleton, wliich " somewhat recalls the shape in OpMoscoIex, though the peristomial plates are entire and trans- versely oval." Ludwig's 'first ventral arm plate' is present at the upper outer corner of the oral slit. The radial shields are long, bar-like, narrow, composed of several overlapping secondary plates soldered together. Again, according fo Ta^ian, in Astrocnida iskUs the digestive cavity has its roof firmly attached to the body wall, but the floor is entirely and the sides partially free. The perihœmal canal is entirely closed, and judging from Lyman's statements, the genital bursa? appear to communicate witli one another so as to form a single compound cavity, regarded by Lyman as the " peri- toneal cavity." I do not know any other genus of the present family with entirely closed periluxîmal canal. My own observations have been made upon Astrochele li/mani and Asleroporpa Jiadracdniha, as representatives of the simple-armed genera of the Gorgonocephcdince. In the first named species, the radial shields are long, narrow, bar-like, almost reach- ing the centre of the disk, composed of several overlapping secondary plates soldered together. 'The genital plates are stout, lying close to the sides of the basal vertebra^. The genital scales are very small, flat," iamella-like, articulated to the genital plates at a short distance from the outer ends of the latter. The basal vertebra) witliin the disk are not covered over by the muscles lying between them and the genital plates. The peristomial plates are entire, very large, roughly pentagonal, with the longest side inwards and the most obtuse angle outwards, much wider than long, closely soldered witli the oral frames, wliich are ratlier MONOCrRArH OF JAPANKSE OPHIUEOIDEA. 63 simple and not very stout. Ludwig's ' ilrst ventral arm plate ' is present at the upper outer corner of the oral slit. The oral plates are very slender and rather long. The interradial attachments of the floor of the gastral cavity extend from the middle of the peri- stomial plates outwards, and the radial attachments from the outer end of the first vertebra3 outwards. The perihaemal canal and genital bursa) communicate with each other, so as to form a single cavity, which is however divided into ten radiating compartments by the radial and interradial attachments of the floor of the gastral cavity. The peritoneal cavity is perfectly divided into five radial compartments, which is entirely filled up by the generative glands. The ventral arm plates are present, being represented by two or three secondary plates. The minute hooks of the arm annulations are comparatively coarse, being composed of from one to three supplementary booklets, besides the main terminal one. In Astci'oporpa hadracantha, the radial shields are entire instead of being composed of several secondary plates, stout and very thick. The interradial attachments of the floor of the gastral cavity ex- tend from the outer end of the peristomial plates outwards. The ventral arm plates are entire, quadrangular, slightly separated by the adradial parts of the b^ir-like lateral plates. The minute hooks are much finer than those of the foregoing species, having only a single supplementary booklet besides the main one, as in many higher genera. The other structures are almost similar to those of the preceding species. Tn higher genera with the arms branched many times, the genital plates lie at the lower lateral sides of the basal vertebrcT, and do not extend inwards to the first vertebra?, as they do in the simple- armed genera. The species of Gorgonocephalus may be divided into three groups on the ])asis of the attachments of the gastral pouches to 64 ART. 2, — H. MATSUMOTO : the basal vertebrae, G. tuberosus representing the first, G. arcticus the second, and G. dolichodadylus the third. In G. tuberosus, the floor of the gastral cavity is almost free from the basal vertebrae, except in the very peripheral parts, where it is attached to the one or two vertebrae just inside the disk border. In G. arcticus, the floor of the gastral cavity is firmly attached to the basal ver- tebrae within the disk, save one or two first vertebrae which are free from the gastral wall. G. cariji and, according to Lyman, also G, cJiilensis belong to this type. In G. doUchodactylus, the attach- ments of the floor of the gastral cavity to the basal vertebrae ex- tend from the second to the sixth vertebrae, so that the two com- partments of the perihaemal canal plus genital bursa3 on either side of a radius communicate together in two places, one lying inwards just above the second vertebra and the other outwards just above the sixth vertebra. In all the three types, the gastral cavity is divided into ten radiating compartments, the radial and interradial gastral pouches, of which the walls again present radiating folds and are thickened by the presence of tlio inner layer of yellowish or brownish glandular cells, the foldings and thickenings being however more prominent in the second and third types than in the first ; and the peritoneal cavity is perfectly divided into five radial compartments, which are entirely fiUed up and obliterated by the very voluminous, strongly folded generative glands. The internal differences appear to me to be correlated with certain external features ; the first type being characterised by the high disk and strongly concave interbrachial ventral surfaces, the second by the low disk and flat interbrachial ventral surfaces, and the third by the very liigh outer ends of the genital slits. Astrodendnim sagaminum is almost similar to the second type of GorgonocephaJus in its internal structure. In Astrocladus anmdatus^ MONOGRAPH OF JAPANESE OPHIUPvOIDEA. 65 the general plan of the internal structure is similar to that of the second type of Gorgonocephalus ; while in Astrodadus conlfcrns, the plan is that of the third type, though the foldings of the gastral pouches, as well as of the generative glands, arc more complex and the layer of t!io glandular cells of the wall of the gastral cavity is extraordinarily thick, so that the cavity itself is much less spacious. In Asfroboa arctos, the internal structure is essen- tially similar to that of the preceding, but there is often a fenestra in each interradial septum between tlie two compartments of the perihaemal canal plus genital bursa? on either side of an interradial line, placing the two in communication with each other. In this species, the inner side of the dorsal surface of the vertebrae is strongly convex, and the outer side correspondingly concave. DÖDEKLEIN has divided the Gorgonocephalidce into two sub- families according to the presence or absence of supplementary plates in the spaces between the oral angles and the interbrachial ventral surfaces. But his statement is in my opinion based on an error, because my Astrotomince evidently have the supplementary plates in question. In Astrothamnus echinacsus, the supplementary plates in question are well developed, and the adorai shields are separated by them from the interbrachial ventral surfaces. Besides, there occurs a not very large supplementary plate among the adorai shields and oral plates. The greater inner parts of the oral plates markedly project ventrally. The oral skeleton is firmly soldered together. The peristomial plates are entire, irregular in shape, with convex dorsal surface. The oral frames are humped dorsally near the outer ends, being much higher than the basal vertebrae. Ludwig's ' first ventral arm plate ' is present at tlie dorsal side of the distal end of each oral slit, being firmly solder- ed with the oral frames. The second to sixth vertebrae are markedly 66 ART. 2. — H. MATSUMOTO : small and narrow, being narrower than the first vertebra and the basal free vertebrae of the arms, and are covered over by the muscles lying between them and the genital plates. The genital plates are very stout, lying close on the sides of the adjacent ver- tebra?. The genital scales are very small, articulated to the outer parts of the genital plates. The radial shields are long, narrow, thick, with humped interior surfaces, nearly reaching the disk centre. The gastral cavity is divided into ten radiating pouches, the roof being firmly attached to the body wall. The floor of the gastral cavity has ten attachments, the radial ones extending from the second vertebra to the disk margin and the interradial ones from the peristomial plate to the same. The perihaemal canal commuicates with the genital bursœ, which are very spacious. The peritoneal cavity is perfectly divided into five radial compart- ments, which are entirely filled up by the generative glands. The ventral arm plates are rhomboidal, much wider than long, separated from one another by the lateral arm plates, which meet each other in the ventral median line. The abradial parts of the lateral plates strongly project ventrally to form a very prominent spine ridge. The minute hooks of the annuli of the arms are compound, being composed of a terminal main booklet and of three to five comb- like supplementary ones. In Astrotoma sobrina, the essential structure is almost similar, but the supplementary plates in the spaces between the oral angles and the interbrachial ventral sur- faces are better developed ; the muscles lying between the basal vertebrae and the genital plates are more massive ; the radial shields are distinctly keeled ventrally ; the radial attachments of the floor of the gastral cavity extend from the first vertebra? to the disk border ; and the minute hooks of the annuli of the arms are composed of from one to three supplementary booklets, besides MONOGEAPH OF JAPANESE OPHIUKOIDEA. 67 the terminal main one. According to Dödeelein, the minute hooks of the annuli of the arms of Astrothorax misakiensis Döderlein have a single supplementary hooklet, besides the terminal main one. The position of the madreporic shield of Astrotoma agassizi Lyivian is almost the same as in Astrospartiis, and such a position of the shield is hardly possible unless the supplementary plates are present in the space inside the interbrachial ventral surface. Key to Japanese genera of Gorgonocephalinœ, A — Arms simple; no supplementary plates in the spaces between the oral angles and the interbrachial ventral sm'faces; annulations of the arms continued onto tlie dorsal side of the disk, so that the latter is concentrically annulated as a whole Asleroporpa. -4^— Arms divided from the base ; supplementary plates present in the spaces between the oral angles and the interbrachial ventral sur- faces; disk not concentrically annulated. a — ^Arm spines present from the very base. b — ^Interbrachial margins of the disk with a row of marginal scales . . Gorgonocephalus. bb — ^Marginal disk scales absent Astrodendrum. aa — Arm spines absent at least within the first bifurcation ; marginal disk scales absent. c — Arm spines present within the second or third bifm*cation Astrocladus. cc — Arm spines absent at least within the fourth bifurcation . . Astroboa. Asteroporpa hadracantha Clark. Asteroporpa liadracantha : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 280, fig. 142. One specimen ; Uji-sliima, Osumi ; 80-90 fathoms. One speci- 68 AET, 2. — H. MATSÜMOTO men; Oniga-sé, Sagami Sea; 150-200 fathoms. One specimen; locality unknown, probably Sagami Sea. Eastern Sea ; 103 fathoms (Clark). Off Omai Zaki, Yenshû Sea ; 34-37 fathoms (Claek). Off Suno Saki, Sagami Sea ; 44-50 fathoms (Clark). The largest one of my specimens measure 7 mm. across the disk and 35 mm. in the length of the arms. In one specimen, some of the super- ficial granules of the disk and arms, both above and below, bear here and there a very long glassy hair. Ophiocrene Bell and the Ophiocj^enoid stage. The curious genus, Ophio- crene Bell, is characterised by the disk being covered mostly by the naked primary plates and by the radial shields being not long and bar-like but short and rounded. I have two veiy small, generically and specifically indeterminable specimens, which are almost similar to Ophiocrene in dorsal view, but differ from it in the occurrence of arm spines even in the very base of the arms. According to Lyman and Grieg, very young specimens of Gorgonocephalus ar oticus Leach {:=G. agassizi) also have the disk essentially similar to that of Ophiocrene. I therefore imagine that, the disk characters of Ophiocrene occur in a certain young stage of many — presumably all — genera of the Gorgojiocephalidce, and Fig. 17. Asteroporp'i hadracantha. x8. a From above, h. From below. MONOGRAPH OP JAPANESE OPHIUROIDEA. 69 propose the name ophiocrcnoid for it. Judging from the presence of a single madreporite and the absence of the arm spines in the basal parts of the arms, Ophiocrene may probably be the ophlo- crenoid stage either of Asfrohoa or of AstrocJialcis ; and if this connection should be proved, Ophiocrene has priority as a generic name. An Ophiocrenoid singe of one of the Gorgonoceplialinœ, Two specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Diameter of disk 2.5 mm. Arms twice divided ; length of the primary shafts 3 mm., of the secondary shafts 2.5 mm., and of the last shafts 1.5 mm. Width of arms at base 0.7 mm. Disk five-lobed, rather high, with strongly concave interbrachial borders, cover- ed by a pavement of naked plates, among which the central, five radiais, five interradiais, as well as the radial shields are most con- spicuous, with convex sur- faces. The secondary plates arc very small and incon- spicuous, and form a system of zones surrounding the primaries. Ventral side of disk covered by a granulated skin, through which the oral and adorai shields can be seen as being entire and without sup- <£>. Fig. 18. An Ophiocrenoid (? Gorgonocephalus or Astrodendnim). x20. a. From above, h. From below. 70 ART. 2. — H. MATSUMOTO : plementaiy plates. Teeth, dental papillae and oral papillae very few in number. Arms covered by a pavement of flattened, polygonal granules, and distinctly annulated by double rows of compound hooks, of which the terminal booklet is smaller than the accessory one. Ventral side of arms covered by a naked skin, through which the ventral and lateral arm plates are visible. Two small, peg-like arm spines occur on the second and outer lateral arm plates. Colour in alcohol white. The smaller one of the two specimens is only 1 mm. across the disk and has the arms divided only once. The disk is entirely covered with the primaries and the radial shields, secondary plates being absent. The arms are not yet annulated by the double rows of compound hooks, but there are from three to five second- ary plates in place of each dorsal arm plates, so that the dorsal view of the arms reminds us of that of Hemieuryale. Key to Japanese species of Gorgonocephalus, A — Arms composed of short shafts ; arm spines almost as long as the corresponding arm joint. a — Teetli and papillae, as weU as arm spines, very acute ; internal structm'e of the first type tuherosus. aa — Teeth and papillae, as weU as arm spines, blunt; internal structure of the second type caryi. AA — Arms composed of long shafts; arm spines shorter than half the corresponding arm joint ; internal structure of the third type dolicliodadylus. Gorgonocephalus ttiberosus Döderlein. Gorgonocephalus tiiberosus : Döderlein, Zool. Anz., XXV, 1902, p. MONOGRAPH OF JAPANESE OPHIUROIDEA. 71 322 ; DÖDERLEIN, Abli. Matli.-Pliys. Kl. K. Bayer. Akacl. Wiss,, Suppl.-Bd. I, 1911, p. 33, PI. II, figs. 1 & 2. Four specimens ; off Misaki, Sagami Sea. Sagami Sea ; 240 m. (Döderlein). The largest one of my specimens is 56 mm. across the disk, which is beset with very numerous rough tubercles. One speci- men (fig. c), measuring 27 mm. across the disk, closely resembles the next species, almost agree- ing with Lyiman's description of the same, except that the apparent granules, which are scattered on the disk, are in reality very fine, rough spinules as ascertained by an examination under the micro- scope. A vertical section of this specimen showed that it had the internal structure of the first type. I am inclined to regard this specimen as possibly a natural hybrid be- ^^Q- 19- Gorgomceplialus tulerosus. a. From above, xl^-. h. From below, xl^. c From tween the typical G. iuberosus above. x2. and the next species, as the two occur almost in the same place. I Wi Gorgonocephaliis caryi (Lyman). Astroplnjton caryi: Lyman, Proc. Boston Soc. Nat. Hist,, VII, 1860, p. 424'^ ; Lyman, 111. Cat. Mus. Comp. ZooL, I, 1865, p. 184. 1) This paper was not seen by me. 72 AP.T. 2.— H. MATSUMÔTO : AstropJiyton stimpsonii : Veeeill, Proc. Boston Soc. Nat. Hist., XII, 1869, p. 888.^) Gorgonocephahi9 caryi : Lyman, K,ep. Challenger, V, 1882, p. 264 ; Clark, Bnll. U. S. Nat. Mus., LXXV, 1911, p. 287. Gorgonocephalus stimpsoni : Lyman, loc. cit., 1882, p. 264. Gorgonoccphaïus japonicus : Dödeelein, Zool. Anz., XXV, 1902, p. 322; DÖDEELEIN, Abh. Matli.-Phys. Kl. K. Bayer. Ak ad. Wiss., Siippl.-Bd. I, 1911, p. 31, PL I, fîgs. 1-3, PI. Vn, figs. l-2c. Gorgonoceplialvs sagaminvs : Doflein, Ostasienfalirt, 1906, i:>. 204, fig. (Non Dödeelein, 1902.) Four specimens ; off Misaki, Sagami Sea. . Sagami Sea ; 150-800 m. (Dödeelein). Off Osé Zaki, Sm^uga Gulf; 63-75 fathoms (Glaek). Eastern Sea; 181-391 fathoms (Claek). Korea Strait ; 66 fathoms (Claek). Korea (Dödeelein). Sea of Japan ; 59-428 fathoms (Claek). Saghalin ; 40-43 fathoms (Claek). Saghalin (Dödeelein). Ochotsk Sea ; 73 fathoms (Claek). Ochotsk Sea (Veeeill). Arctic Sea. Bering Sea. Alaska to California. One of my specimens is evidently the japonicus type, the surfaces just inside the interbrachial ventral surfaces, as well as the ventral side of the arms being entirely free of granules. The lower margins of both sides of the arms are covered with irregularly polyonal plates, without granules, as shown in Dödee- lein's fig. 2a, PL VII. Dödeelein has not found the canj'i type in Japanese waters. But my two specimens agree well with Ly]vian's description, the granulations of the interbrachial ventral surfaces distinctly extending as far as the base of the oral angles and the outer end of the oral slits. The ventral side of the arms is rather sparsely granulated, the granules being however finer 1) This paper -was not seen by me. MONOGRAPH OF JAPANESE OPHILiROIDEA. 73 than those of the outer parts. The lateral sides of the arms are entirely covered with fine granules, as shown in Dödeelein's fig. la, PI. VII. In the arm coverings, these two specimens are therefore quite similar to Dödeelein's ''japonicus var." The disk is rather sparsely and uniformly granulated. The internal structure is of the second type in aU the four specimens. Govgonoceplialus dolicJiodactyliis Döderlein. Gorgonoceplialiis dolicJiodactylus : Döderlein, Abb. Math.-Pliys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 34, PL I, figs. 4 & 5, PL VII, figs. 3-4&. Two specimens ; off" Misaki, Sagami Sea. Sagami Sea ; 150- 200 m. (Döderlein). One of these specimens is very ab- normal, all the pairs of the radial shields being fused into one almost along the entire adradial border, though remaining separate to- wards the centre, so that there is here a Fig. 20. Gorgonocephalus dolichodadylus. bifurcation. above. &. From beloAv. xli. a. From Asfrodendruni sagaminiini (Döderlein). Gorgonocephalus sagaminus : Döderlein, ZooL Anz., XXV, 1902, p. 321 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 292. 74 AET. 2. — H. MATSUMOTO : Astrodendrum sagaminum : Döderlein, Abli. Math.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd., I, 1911, p. 38, PI. H, figs. 3-5, PI. Yn, fig. 8; BoMFOED, Bee. Ind. Mus., IX, Pt. 4, 1913, p. 220. Nnmerous specimens ; off Misaki, Sagami Sea; 100-200 a fathoms. ,^w^''*^ Ose Zaki, Sm-uga Gulf; 63- ASW ■^im^^^^' 75 fathoms (Claek). Eastern Sea ; 95 fathoms (Clark). Sea of Japan ; 59-172 fathoms (Claek). Sagami Sea (Dödeelein). vj^V Indian Ocean (Bomfoed). ÂJ* ',s^^^^^m' • Ar* 'w^ ^%*'* ^^-N Key to Japanese species and '^'^^ wl y »'f^ varieties of Astrocladus, '^ j w «"^ Fig. 21. Astrodeiidrum sagaminum. x3. ^-Disk covered by an apparently ^^ p^^^ ^^^^^ ^ ^^^^ ^^,^^ smooth skin, in which how- ever very fine, insignificant, close-set granules are contained ; internal structure as in the second type of GorgonocepJialus annulât lis. A A — Dish covered by a pavement of granules, which have acute, spiny tips ; internal structure as in the third type of GorgonocepJialus .... coniferus. a — Disk and arms free of tubercles, or a conical, blunt tubercle may occur on each radial rib near the outer end. h — Colour variegated, with whitish ground colour and purpHsh brown patches of various shape and size var. pardalis. hh — Colour simple, deep pm-pHsh brown or purplish black typical coniferus. aa — Disk and arms beset with numerous hemispherical, or more or less conical, tubercles; colour variegated, or simply deep pm'pHsh brown or purplish black var. doßeini. MONOGRAPH OF JAPANESE OPHIUROroEA, 75 Astrocladus annulatus Matsumoto. Astrodadus annulatus : Matsumoto, Proc. Acad. Nat. Sei., 1915, p. 56. One specimen ; off Misaki, Sagami Sea. Diameter of disk 22 mm. Distance from the centre of the disk to the interbrachial margin 8.5 mm. From the outer end of the oral sht to the 1'^- bifurcation 11mm. From the disk margin to the l"'" bifurca- tion 1.5 mm. From the 1'*' bifurcation to the 2°'^- .... 7 mm. From the 2"'- to the 3"''' 8.5 mm.-6.5 mm. I From the S'"'^- to the 4'^- 5.5 mm.-9 mm. I From the 4"'- to the 5'^ 9 mm.-4 mm. 1 From the 5"' to the 6"'- 4 mm.-9 mm. From the 6"'- to the 7* 10 mm.-4 mm. From the 7"'- to the 8"' 4 mm.-7.5 mm. From the 8'"- to the 9"" 7 mm.-3.5 mm. From the 9''- to the 10*"- 3.5 mm.-8 mm. From the 10*'^- to the IT"- 8 mm.-3.5 mm. From the 11"^- to the 12"^- 3 mm. -7 mm. From the 12'"- to the 13'"- 9 mm.-3 mm. I From the 13'''- to the 14'"- .. .. . . 2.5 mm.-5 mm. From the 14*'' to the 15* • 5 mm.-2.5 mm. From the 15''- to the 16*^ 2 mm.-3.5 mm. From the le*'' to the 17*^- 3.5 mm.-2 mm. From the 17"' to the 18*''- 1.5 mm.-3.5 mm. From the 18*''- to the 19*- 3 mm.-1.5 mm. From the 19"'' to the end 1 mm.-2 mm. 76 AET. 2. — H. MATSUMOTO Total length of the arm 12G mm. Width of the ventral side of the arm base within the disk .... .... .. ' 4.5 mm. Disk five-lobed, with concave interbrachial borders, covered by a thick skin, which is apparently smooth, but contains very fine and close-set grannies of microscopic size. On the radial ribs, these granules are very flattened and smooth, and coarser, being visible even to the naked eye. Several smooth, hemispherical tubercles are scat- tered on the disk. Eadial ribs gently raised and forming rounded ridges, with rather in- distinct outlines, not quite reaching the centre ; their back, is marked with concentrically arranged swellings, corresponding to the imbricating, soldered plates, of which the radial shields are composed. The ventral surface of the disk appears very smooth to the naked eye. Genital slits not very large. Madreporic shield single, at the inner corner of one of the interbrachial ventral sur- faces, small, transversely oval. The teeth and dental papillœ are conical and rather stout. The oral papilla? and lower dental papill99 are smaller and very short ; their number for each oral angle is not very great. Arms slender and branched, with distinction of trunk and Fig. 22. Astrocladus annuluius. xl^. a. From above. h. From below. MONOGKAPH OF JAPANESE OPHIÜEOIDEA. 77 lateral branch even at the basal region, covered on the upper side by a finely and closely granulated skin, with several scattered, smooth, hemispherical tubercles on the more proximal shafts, dis- tinctly annulated with hook-bearing segments throughout. Ventral surface of the arm entirely smooth to the naked eye. The arm spines which are present beyond the first bifurcation, are very fine, and three or four in number for each tentacle pore. Colour in alcohol : disk mottled, arms annulated with yellow- ish and grayish brown. This species can be easily distinguished from the other species of Astrocladus by the entirely smooth disk covering and by the arms being distinctly annulated with hook-bearing segments even at the very basal region. Astrocladus coniferus (Döderlein). Astrocladus coniferus var. pardcdis (Döderlein). Astrocladus coniferus var. dofleini (Döderlein). Astroplujton x^ardalis : Döderlein, Zool. Anz., XXV, 1902, p. 323 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 293. Astropliyton coniferum : Döderlein, loc. cit., p. 325. Astrocladus dofleini: Döderlein, Schultze — Zool. Ergebn., IV, 1910, p. 256^); DÖDERLEIN, Abh. Math.-Phys. Kl. K. Bayer. Akad. Wiss, Suppl.- Bd. I, 1911, p. 41, fig. 9, PI. n, fig. 6, PL m, figs. 1-4, PL IV, figs. 15- 15&; BoMFORD, Kec. Ind. Mus., IX, Pt. 4, 1913, p. 200, PL XIH, fig. 1. Astropliyton cornutum : Clark, Bull. U._ S. Nat. Mus., LXXV, 1911, p. 292. (Non Kœhler, 1909.) Astrocladus coniferus: Döderlein, loc. cit., 1911, p. 46 & 76, PL IE, figs. 7 & la, PL IV, figs. 1-3«, PL VI, figs. 5-6a & 16. Numerous specimens ; Misaki. Several specimens ; off Moroiso, Misaki ; 10 fathoms. 1) This paper wn?, not seen by me. 78 ART. 2. — H. MATSUMOTO Sagami Sea. Suruga Gulf; 108 fathoms (Claek). Off Omai Zaki, Yenshii Sea, 34-36 fathoms (Claek). Kagoshima (Dödeelein). Colnett Strait; 83-84 fathoms (Claek). Eastern Sea; 103-152 fathoms (Claek). Fusan, Korea (Claek). Wladiwostok (Dödeelein). ? Off Agattu Is. ; 482 fathoms (Claek). Indian Ocean. I fail to find any marked distinction between the coniferus and the dofleini type. There are many specimens which must be referred to the coniferus type in most characters, but are beset with a few or many large hemispherical tubercles on the dorsal side of the arms, as observed by Dödeelein himself, but there are others which gradually pass over into the dofleini type. The large tubercle on the back of each radial shield near its outer end, present in most speci- mens of the coniferus and in some of the 2:)ardaUs type, is also often present in smaller and moderately large speci- mens of the dofleini type. Such a tubercle is indicated on some of the radial shields in Dödeelein's fig. 5, PL IV. This tubercle later becomes quit>«-,Q indistinguishable, owing to the increase in number of similar «/v Fig. 23. Astrodadus coniferus. a. Erom above. x|. 1). From below. x|. c. From below. x2. MONOGRAPH OP JAPANESE OPHIUROIDEA. 79 tubercles on the entire dorsal surface of the disk. In distinguish- ing A. dofleini from the coniferus and pardalis type, Döderlein seems to lay weight upon the fact that, the granules of the arm bases of the former have no acute tips, but are smooth. But, one of my specimens, measuring 55 mm. across the disk, has very abundant hemispherical tubercles on the arms and easily recognizable acute granules on the arm bases, the latter being distributed as far out as the fifth bifurcation. The colour of the same specimen is simply deep purplish black, and the disk coverings are almost similar to those of the coniferus type, though the conical tubercles on the radial ribs near the outer ends are not present. Key to Japanese species of Astrohoa. A — Madreporic shield lying at the inner angle of the interbrachial ventral surface. a — Annulations of double rows of hook-bearing granules distinct even in the stout proximal parts of arms ; colour purphsli black . . nigra, aa — Annulations of double rows of hook-bearing granules absent in the stout proximal parts of arms; colour dark grayish brown. . .arctos. AA — Madreporic sliield lying on the hard part just inside the inter- brachial ventral surface, being almost free from the latter ; colour light pinkish brown glohifera. Astroboa nigra Döderlein. Astropliyton davatmn : Pfeffer, Mitteil. Naturh. Mus. Hamburg, XIII, 1896, p. 48.^) (Non Lyman.) Astroboa nigra : Döderlein, Abh. Math.-Phys. 111. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 83, PI. IX, figs. 9 & 9a. 1) This paper was not seen by me. 80 ART. 2. — H. MATSUMOTO: Hirado Strait ; 72 m. (Döderlein). Zanzibar (Dödeelein). Astrohoa arctos Matsumoto. Astrohoa arctos : Matsumoto, Proc. Acad. Nat. Sei., 1915, p. 57. Two specimens ; off Moroiso, Misaki ; 5-10 fathoms. Four specimens ; off Misaki. Diameter of disk 65 mm. Distance from the centre of tlie disk to the interbrachial margin. . . 25 mm. From the outer end of the oral slit to the I''*- bifurcation 32 mm. From the I'''- bifurcation to the 2"'^- 10 mm. From the disk margin to the 2"'^- bifurcation 4 mm. From the 2°'^- to the S'*^- .... 12 mm.-16 mm 14 mm.-13 mm. I I From the 3"'- to the 4*'''- 18 mm.-8 mm. ..11 mm.-15 mm. I I From the 4"^' to the 5*'' .... 10 mm.-18 mm. , 16 mm.-ll mm. 1 1 From the 5''' to the 6"'- 17 mm.-9 mm.. .10 mm.-18 mm. I I From the 6"'- to the 7"'' .... 9 mm.-16 mm 14 mm.-ll mm. 1 1 From the 7"'- to the 8*- 15 mm.-8 mm. . . 9 mm.-17 mm. I I From the 8*''- to the 9*' .... 8 mm.-16 mm 14 mm.-ll mm. 1 I From the 9*- to the 10"^- 16 mm.-8 mm. . . 8 mm.-14 mm. I I From the 10"'- to the 11*'' ■ , . 8 mm.-16 mm. , 13 mm.-8 mm. I I From the ll'''- to the 12'''- 15 mm.-6 mm. . . 7 mm.-20 mm. I I From the 12*''- to the 13"'' . . 6 mm.-17 mm^ 11 mm.-7 mm. I I From the 13"' to the 14*''- 15 mm.-6 mm. , . 7 mm.-ll mm. I I From the 14"'' to the 15'''- . . 6 mm.-lO mm 10 mm.-5 mm. 1 I From the 15"'' to the 16*^ 14 mm.-6 mm. . . 5 mm.-9 mm. i I From the 16*''- to the 17* . . 6 mm.-ll mm 11 mm.-5 mm. I I From the 17"'- to the 18"'- 14 mm.-5 mm. . . 6 mm.-9 mm. I I From the 18*''- to the 19"'- . . 4 mm.-12 mm 14 mm.-5 mm. I 1 From the 19*''' to the 20*''' 16 mm.-5 mm. . . 6 mm.-lO mm. MONOGRAPH OF J-KPANESE OPHIUROIDEA. 81 Fi-om the 20**^- to the 21"'- . . 5 mm.-12 mm .9 mm.-4 mm I I From the 2V' to the 22"^" 11 mm.-5 mm. . . 4 mm.- 9 mm. 1 I From the 22"'^- to the 23'^'^ . . 4 mm.-lO mm 9 mm.-4 mm. I I From the 23''''- to tlie 24*''- 8 mm.-4 mm. . .4 mm.- 9 mm. I I Fi-om the 24"'- to the 25^'' . . 4 mm.-lO mm 9 mm.-4 mm. I I From the 25'''- to the 26'" . . 11 mm.^ mm. . . 3 mm.- 6 mm. I I From the 26* to the 27'" . . 4 mm.-li mm 7 mm.-3 mm, I ! From the 27'*'- to the 28'" 9 mm. -3 mm. 3 mm.- 6 mm. I I From the 28'"- to the 29'" . . 5 mm.- 9 mm 4 mm.-3 mm. ' I From the 29'"- to the 30'"- 7 mm. -4 mm. . . 2 mm.- 4 mm. I I From the 30'"- to the 31«' . . 3 mm.- 8 mm 3 mm.-2 mm. I I From the 31^'- to the 32°*^ 6 mm. -4 mm. . . 2 mm.- 3 mm. From the 32"'*- to the 33'^'' (or | | to the end) 3 mm.- 6 mm 2 mm.-l mm. I From the 33'"'^- tu tlie 34'"- 6 mm.-4 mm. I From the 34'" to the 35'" . . 4 mm.- 5 mm. 1 From the 35'" to the 36'"- 3 mm.-2 mm. 1 From the 36'"- to the 37'"- . . 2 mm.- 3 mm. I From the 37'"' to the end 2 mm.-l mm. Total length of the arm 421 mm 324 mm. Width of the ventral sm:^ace of the arm within the disk 17 mm. Width of the shaft between the 1^ and 2^'^- bifm:cations 10 mmi Width of the main (adradial) shaft between the 2"*^ and 3'^'' bifur- cations 7 mm. Disk decagonal, with concave interbrachial and brachial borders, the former being longer and more concave than the latter ; very- convex, but with more or less depressed central region, covered by a thick skin, which is chagreened by the presence of very fine, close-set granules. The granules are smooth, not acute, irregular in size when viewed under the microscope, the coarser ones being more numerous on the radial ribs than in the intercostal spaces. 82 ART. 2.— H. MATSUMOTO : Fig. 24. Aströbon arcto.t. b. From below. X li. «. From Mbove. The radial ribs are bar-like, narrow, widest at the outer end, suddenly narrowed for a very short distance, then uniformly tapered inwards, nearly reaching to the cen- tre. Interbrachial ventral surfaces covered by a thick, apparently smooth skin, which contains very fine granules. Adradial border of the genital slits closely spinulated. The madreporic shield, situated at the inner angle of the interbrachial ventral surface, is more or less semilunar, with a semicircular inner side and a distinctly notched outer side, the lateral angles being rounded. The spaces inside the interbrachial ventral surfaces are apparently smooth, but in reahty closely covered with very fine granules of microscopic size, the granules being rather coarse and distinct at the oral angles. Teeth and papillae very numerous. The oral and dental papillae are rather small, spiniform, not very acute, while the teeth are much larger and longer than the papillae, and are distinctly spatulate and flattened at the tip, the upper teeth being larger and more pronouncedly spatulate than the lower. The two main stems of each arm are not equally developed, but one is longer, stouter, and richer in bifurcations than the »ther, as shown in the preceding table of dimensions. Dorsal and lateral sides of arms covered by a thick skin, which is very MONOGRAPH OF JAPANESE OPHIUROIDEA. 83 finely and closely granulated, the grannies being very irregular in size and roughly distinguishable into two kinds. The fi.ner granules are entirely covered over by the skin, and are flat and irregularly polygonal, forming together a sort of mosaic ; while the coarser granules, which are coarser than any granules of the disk, are hemispherical and tubercle-like, and are uniformly scattered. Ventral surface of arms apparently smooth,' but really covered by a mosaic of flat and irregularly polygonal granules of microscopic size. The first pair of tentacle pores are distinct and open in slight depres- sions. The second are often visible. The following three or four pairs are entirely absent, those beyond being again distinct. The arm spines are entirely absent in the proximal joints, but are present from the fourth or fifth bifurcation outwards. They are exceedingly minute and granule-like, two to four of them occurring for each tentacle pore. The double rows of hook-bearing granules are present only in very fine twigs, the main stems within the fourteenth or fifteenth bifurcation being free of them. The shaft between the first and second bifurcations usually contains four arm joints. The outer shafts are composed of from six to eight, usually seven, joints. Colour in alcohol, as well as in a dry state ; dark grayish brown above, and dark yellowish brown below. ... Of the five known species of Astrohoa, A. clavata (Lyman) is distinguished from the others by the spiny granules of the disk and arms, and A. globifera (Döderlein) by the position of the madreporic shield. A. mida (Lyivian) and A. nigra Dödeelein have distinct annulations of hook-bearing granules on the arms, while A. ernce Döderlein has no such annulation on the greater proximal parts of the arms. This species is therefore very near to the last named but differs from it in the less distinct and much finer 84 ART. 2. — H. MATSUMOTO : grannies of the disk, and in the less number of arm joints con- tained in a shaft. In the last characters, A. arctos rather resem- bles A. nigra from Zanzibar and Hirado Strait, Japan. This species is common in the shallow waters aronnd Misaki,. and occurs in the same localities with the foregoing species. Astroboa glohifera (Döderlein). AstropJiyton glohi/ernm : Döderlein, Zool. Auz., XXV, 1902, p. 324. Astroboa glohifera: Döderlein, Abh. Matli.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 51, PI. n, figs. 8-9 ; PI. VU, figs. 7 & la. One specimen ; off Kotsu-jima, Izu. Sagami 8ea ; 150-200 m. (Döderlein). Key to genera of Astroto^ninœ, A — Interannuli of the arm covered with two rows of coarse granules ; arms simple. a — Disk more or less closely beset with thorny tubercles or spiny granules Astrothamnus. aa — Disk covered with rounded plates, which are surrounded by belts of granules. I — Disk plates irregular in size ; two or three arm spines Astrothrombns. hh — Disk plates regular in size ; three to seven arm ^^ines., Astrothorax^ AA — Interannuli of the arm covered with many rows of fine granules ; disk also covered with fine granules. c — Arms simple Asfrotonia^ cc — Arms divided a few times near the tip Aatroclon. Astrothamnus Matsumoto, 1915 Disk divided into ten radiating lobes by radial and interradial furrows, closely covered with coarse, thick granules, which are MONOGRAPH OF JAPANESE OPHIUJiOIDEA. 85 more or less stumpy and usually have spiny tips. Interradial ■ventral surfaces strongly concave, each with two large, nearly parallel genital slits. A single madreporic shield is present at the inner border of an interbracliial ventral surface. Teeth and dental papillae similar, spiniform, forming clusters at the apices of jaws. Oral papillae absent, so that the sides of the oral slits are naked. Arms simple, covered by a pavement of gi-anules, distinctly an- nulated by zones of densely set, minute, compound hooks, each of which consists of one main and several supplementary booklets. Three to five arm spines, peg-like, usually with rough tips, serving as tentacle scales. This genus includes Astrotoma hellator Kœhler, 1904, vecors Kœhler, 1904, and rigens Kœhlee, 1910, besides the genotype, Astrothamnus echinaceus. They may be distinguished as follows. Key to species of Astrothatnnus. A — Brachial ventral surfaces smooth ; spaces just inside the interbrachial ventral surfaces, as Avell as oral angles, beset with slender spines ; three arm spines. « — Basal arm spines unusually long and stout, longer than the cor- responding arm joints rigens, aa — Basal arm spines not unusually long and stout, nearly as long as the corresponding arm joints vecors. A A — Brachial ventral surfaces, spaces just inside the interbrachial ventral surfaces, as well as oral angles, beset with coarse, stumpy granules. h — Three arm spines not very small, nearly as long as the correspond- ing arm joints echinaceus. bb — Five arm spines Aery small, hardly half as long as the correspond- ing arm joints bellator. The distribution of Astrothamnus is very interesting, for, as 86 ART. 2. — H. MATÖUMOTO shown in the foregoing key, there are two specific groups in tliis genus, each represented by a Malaysian species. The Arabian species, A. rigens^ is nearly allied to the Malaysian A. vecors, and the Japanese type, A. echinaceus, to the Malaysian A. hellator. Astrothamnus echinaceus Matsumoto. AstrotJiamtins echinaceus ; Matsumoto, Proc. Acad. Nat. Sei. Philadel- phia. 1915, x>. 55. Diameter of disk 22 mm. Length of arms 140 mm. Width of arms at base 4 mm. Disk distinctly five-lobed, with five interradial furrows, each lobe being again divided into two secondary lobes by the radial furrow. Radial ribs very prominent, large, occupying almost the whole dorsal surface of disk, leaving between them only ten . narrow furrows radiating from the centi'e ; closely covered with rather large stumpy tubercles with thorny crowns, between which lie thick, irregularly polygonal plates. Inter- brachial ventral surfaces strongly concave, closely covered with stumpy tuber- cles terminating with one or a few thorny points. Genital slits rather large, more or less parallel. Mad- reporic shield distinct, small» Fig. 2S. Astrothamnus echhmcem. x2^. a. irrCgular in OUtHnC. The From above, b. From below, c. Side view of three arm joints near disk. . area inside the iuterbrachial MONOGRAPH OF J.VrANESE OPHIUßOlDEA. 87 ventral surfaces, as well as the brachial spaces', are also closely covered with tubercles similar in form to those of the interbrachial ventral surfaces. Oral angles covered with close-set, sharp, conical tubercles,, which become, towards the mouth, somewhat indistin- guishable from the dental papillae. Dental papillae and teeth similar, conical and very acute. No oral papillae. Arms rather slender, uniformly tapered, with alternating annuli of two forms, one with two rows of granules entirely covered over with very densely set minute hooks, the other with two irregular rows of smooth, naked granules. Ventro-laterally on either side of the arms, in line with the annuli formed by the smooth granules, there is a series of large, round, naked plates,. In the arm bases, the hook-covered annuli are usually broken at the dorsal median line by conical granules terminating with one or a few thorny points. Ventral surface of the arms with rather well spaced tubercles, which are conical or terminate with one or a few thorny points ; these tubercles become rounded "äiid smooth distally. The first and second tentacle pores without arm äpineä, the third with one or two spines, the fourth with two or three, and the rest with three. The arm spines of the basal pores are somewhat indistinguishable from the conical or thorned tubercles» but the rest are peg-like, nearly as long as the corresponding arm joints, and bear two or three denticles at the end. The oral tentacle pores and the first and second tentacle pores open by means of short tubes bearing a few spinules. : Colour in alcohol dull grayish purple. Astrofhoraoc misakiensis Döderlein. Astrothorox misakiensis : Döderlein, Abh. Matli.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 24, PI. VI, figs. 2-26, PI. VII, figs. 12 & 14&. 88 ART. 2. — H. MATSÜMOTO Sagami Sea (Döderlein). Astrotoma sobrina Matsümoto, Astrotoma murrayi : Döderlein (non Lyman, 1879), Abb. Matli.-Phys. Kl. K. Bayer. Akad. Wiss., Suppl.-Bd. I, 1911, p. 23, fig. 1, PI. VI, figs. 1 & la, PI. VII, figs. 14-146. Five specimens ; off Misaki, Sagami Sea. Diameter of disk 34 mm. Length of arms 200 mm. Width erf arms at base 7 mm. Disk five-lobed, flat, rather sunken towards the centre, covered with very fine, smooth, closely soldered granules and many small, oval plates, which are sunken below the level of the granules ; there are also on the disk numerous, large, smooth, spherical or ^ghtly cylindrical tubercles with rounded ends. Radial ribs almost indistinguishable. Inter- brachial ventral surface covered by a thick, leather-like skin, which becomes gi-anulated on drying. Genital slits rather large and divergent. The madreporic shield is small, but very distinct, more or less oval, vertical, lying at the inner corner of one of the interbrachial ventral surfaces. The area inside the ventral interbrachial surfaces, as well as the brachial spaces, are covered with thick, irregular, pavement-hke plates and smooth, scattered tubercles, which are F*ig. 26. Astrotomn sohrlnn. This order is very near to the Phrynopliiurida, especially the Ophiomi/xldœ, but differs fi-om them essentially in the entirely MONOGIiAPH OF .lArANESE OPHIUROIDEA. :91 protected doi'sal side of the arms. Further, there are two types of primitive ophiurans differing in the embryonal condition of the vertebrae, one being the Phrynophiuridan type and the other the Laemophiuridan. In the Phrynophiuridan type, the distal vertebrae are divided into halves by a single fusiform pore, the middle parts of the halves being widely separated from each other. These vertebra? are proved to be present in Ophioheliis, Ophiogerofu Astrogeron, Ophiosciasma, Opliiohyalus, Ophiomyxa, Ophiostiha, &c» In the Laemophiuridan type, the distal vertebrae are divided into halves by a series of small pores, the halves being connected with eacli other by a series of bridges between every two pores* These vertebra3 are present in Mlcrophiura, Ophiothamnus, Ophio- logimus, Amphiactis, &g. In AsfrophJura, the vertebrae of the free arm joints are divided into halves by a moniliform pore, the halves being closely set against each other, so that the condition, though rather near to the Laemophiuridan type, is intermediate between the two types. Key to the families of Lœniophiurida, A — Disk and arms of slender build, the disk scales or plates, as well as arm plates, being not very stout ; the genital plate and scale on either side of a radius articulate with each other, instead of being soldered together; vertebrae not very stout, often divided into halves OpMacanthidœ. A A — Disk and arms of very heavy build, the disk and arm plates being very stout ; the genital plate and scale on either side of a radius are firmly soldered together ; vertebrae very stout, always entire, thougli til« ventral median groove is very prominent . . HemieuryolidcB . V2 ART. 2. — H. MATSUMOTO : Family 1. Ophiacanthidse (Pekkiee, 1891) mihi, 1915. Disk covered by a soft skin or witli fine scales, and usually beset with granules, spines or stumpy tubercles. The radial shield and genital plate articulate with each other by means of a trans- verse ridge or a simple face on both the plates. The genital plate and scale on either side of each radius articulate with each other, and are not soldered together. Peristomial plates entire, or rarely triple, very large. Oral frames entire, without well developed lateral wings. Teeth and oral papillae always present, spiniform as a rule. Dental papillae rarely present, similar to the oral papillae. Arms usually knotted, the spine ridges of the lateral arm plates being very prominent ; inserted vei-y often laterally, but sometimes ventrally, to the disk. Dorsal and ventral arm plates very small, while the lateral arm plates are very well developed, covering most parts of the arms. Numerous arm spines, variable in number in successive arm joints, long, flagellate, often hyaline and serrate. A^ertebrae slender, often incompletely di\ided into halves by a series of pores. Vertebral articulation zygospondyline or streptospondyline. Upper and lower nmscular fossae equally large. This family includes thirty-three genera, which may be artificially grouped as follows : A. Tentacle pores very large and open. I. Disk extraordinarily bulged up, covered by a naked skin or with fine scales ; radial shields absent. Ophiotholia Lyman, 1880. Ophiomyces Lyman, 1869. II. Disk not very high, covered by a naked skin or mere- ly with fine scales ; radial shields small or rudimentary. MONOGEAPH OF JAPANESE 0PHIÜR01DEA. 93 Ophiocimbium Lyman, 1880. Ophiologimus Clark, 19.11. Amphipsila Veerill, 1899. Ophiophrura Clark, 1911. , Ophiotoma Lyäian, 1883. ; Ophiohlenna Lütken, 1859. III. Disk beset with grannies or spines ; radial shields small or rudimentary. Opliiopora Verrill, 1899.'^ Ophiotrema Kœhler, 1907.*^ Ophiamhix Lyman, 1880.^^ Ophiomedea Kœhler, 1906, Ophiopristis Verrill, 1899. Ophioprium Clark, 1915. B. Tentacle pores smaU and inconspicaons. IV. Radial shields small, short. Mlcrophlura Mortensen, 1910. Ophiolimna Verrill, 1899. Ophiomitrella Verrill, 1899. Ophîophthalmus, nov. Ophientrema Verrill, 1899. Ophioscalus Verrill, 1899. Ophiocopa Lyman, 1893. Ophiochondrella Verrill, 1899. Ophiosemnotes, nov. 1) These genera have been stated by Clark, 1915, to be congeneric with Ophiotoma, notwith- standing that they differ from it in the covering of the disk. It is of course true that these three genera resemble one another in the large tentacle jiores. But this character is also common to all the genera belonging to the group A. So that I am obliged to disagree with him. 2) This genus is very doubtful in position. I provisionally refer it here, merely because the tentacle pores and scales remind us of those of Ophiophrura and Ophiotrema. 94 ART. 2. H. MATSÜMOTO : V. Radial shields long, narrow, bar-like. Ophiacantha Müüler à Tröschel, 1842. Ophialcœa Verrill, 1899. Ophiacanthella Verrill, 1899. Ophiolebes Lyman, 1878. VI. Radial shields large, long, wide. OpJdothamnus Lyman, 1860 (— OplHoleda Kœhler, 1906). Ophiurothamnus, nov. Ophiomytis Kœhler, 1904. Ophioplinthaca Verrill, 1899. Ophiomitra Lyman, 1869. Ophiocamnx Lyman, 1878. As the present family includes both those forms with only horizontally flexible arms and those with more or less coiled arms, two types of internal structures are also distinguishable roughly. The first type, found in those forms with only horizontally flexible arms, is characterised by the very thin, delicate, sometimes divided peristomial plates, the slender oral plates and frames, the very slender vertebrae, the very thin wings of the basal vertebrae, the more or less divided distal vertebrae and the zygospondyline vertebral articulation. The second type, found in those with more or less coiled arms, is characterised by the very thick, always entire peristomial plates, the more or less stout oral plates and frames, the more or less short, stout, always entire vertebne, the more or less thick wings of the basal vertebrae and the usually strepto- spondyline vertebral articulation. l^he internal structures of Ophiothamnus venustiis Matsumoto is almost perfectly similar to those of 0. vlcarius Lyman. The peristomial plates are very large, thin and triple, the two paired secondary plates forming an out- MONOGRAPH OF JAPANESE OPHIDROIDEÄ. 95 wardly open angle, which is occupied by the unpaired median one. The oral plates and frames are short and slender. The genital plates are long, more or less club-shaped, lying closely parallel in pairs, just above the arm base. The genital scales are absent. The genital bursas ate very rudimentary, being represented merely by the creases between the interbrachial ventral surfaces and the arm bases. The generative glands are enclosed in a membranous sac, the waU of which contains fine, thin, transparent scales, when viewed under the microscope, just as is stated by Moeïensen to be the case in Ophiopus arcticus. The vertebrae are very slender, and those of the distal arm joints are imperfectly divided into halves by a series of pores, just as in Mlcrophlura according to Mortensen. The last named genus is stated to lack the genital plates and scales and to have entire or divided peristomial plates, besides an additional plate, which is perforated by a pore just between the peristomial plates and the oral shield. I imagine that, the perforated plate just referred to may correspond to the unpaired secondary plate of the peristomial system, because I know that in certain genera the unpaired secondary plate has a half pore on its outer border. In OphioUmna antarctlca (Lyman) and 0. papillata (Clark), the peristomial plates are large, wide, short, thin, imperfectly divided, with soldered halves, the oral plates and frames are short and slender, and the genital plates and scales and genital bursae are normal. In Opliiolocjimus hcxactis Clark, the internal structures are essentially similar to those of OphioUmna, which has also imperfectly divided peristomial plates. The radial ' shields though externally invisible, are present and short and rounded. Also in this species I was able to prove that the distal vertebrae are imperfectly divided by a series of pores. According to Lyman, Ophiomyces 96 ART. 2. — H. MATSÜMOTO. friitectosus Lyman has !no peristomial plates, but the oral plates and frames are of an elegant shape and cnrionsly twisted ; the genital plates are thin, wide, long, and curved over the dorsal side of the arm base ; no genital scales ; the basal vertebra?, have large and thin wings without marginal grooves, and the vertebral articulation is very peculiar with a large articular umbo and ne articular peg. Certain species of Ophiacantha appear to stand at the very base of the second type, which is found in those forms with the arms more or less capable of coiling vertically. In Ophiacantha bidentata (Linne), the peristomial plates are very large, simple, rather thick, the oral plates and frames are more or less stout,. the vertebrae are comparatively stout, and the vertebral articulation is perfectly streptospondyline, the articular peg being absent. On the contrary, Ophiacantha cuspidata Lyman is stated by the author to have the zygospondyline vertebral articulation, with a well formed articular peg. In Ophiolebes tuherosa Matsumoto, the peri- stomial plates are very large, wide, thick, firmly fixed to the oral frames, which are also very stout, and bear each two small supplementary plates on the outer border. The vertebrae are short and stout, with decidedly thick wings. The vertebral articulation is of course streptospondyline. The genital plates and scales are pecuharly undulated. The radial shields are narrow and bar-like. The internal structures of Ophiosemnotes œdidisca (Claek) much resemble those of the preceding, save that the radial shields are wide, rounded and joined in pairs in internal view. According to Lyman, Ophiocajnax hystrix Lyman has very stout oral plates and frames, to which the peristomial plates are perfectly soldered, very stout and short vertebra? with very thick wings and a perfectly streptospondyline vertebral articulation. The vertebrae of Ophio- MONOGEAPH OF JAPANESE OPHIUROIDEA. 97 cliondreUa squamosa (Lyman) are quite like those of Ophiolehes. Taking all the characters into consideration, Ophiolehes, Ophiosemnotesy OpMochondreUa, Ophiocamax, &c. appear to me to represent one of the two types of the Ophîacanihidœ. Key to Japanese genera of Oiyhiacanthidce, A — Tentacle pores very large and open ; oral shields entirely separated from the first lateral arm plates by the adorai sliields. a — Disk extraordinarily bulged up, covered with fine scales ; radial shields absent ; numerous large, flat, spatulate oral papillae ; numerous flattened arm spines ; tentacle scales spatulate Ophiomyces. aa — Disk not very high. h — Disk finely imbricated, entirely free of granules or spines ; tentacle scales present, c — Outer oral papilla) not especially long ; ventral arm plates longer than wide ; three arm spines ; two very small, leaf -like tentacle scales OpJdologimus, cc — Outermost three oral papillaa very long, slender, spiniform, arising from the adorai shields; ventral arm plates wider than long; four arm spines ; three or four spiniform tentacle scales .... Op>Iiiophrura, lib — Disk closely beset with granules or spines. d — Dorsal arm i)lates entire; no tentacle scales. .......... . Ophiopora. dd — Dorsal arm ^jlates absent, dorsal side of arms, as well as disk, being closely beset with spines ; three or four spiniform tentacle scales . . Ophiambix A A — Tentacle pores small and inconspicuous. e — ^Ajm plates and arm spines not covered over by a cereous skin. / — Radial shields small or moderate. g — Radial shields short, rounded or triangular. h. — ^Disk closely covered with granules ; oral shields separated from the first lateral arm plates by the adorai shields ; oral angles beset with granules ; outermost oral papilla large and operculiforra . . OphioUrmia. 98 AET. 2. — H. MATSUMOTO. lih — Disk scales visible, sparsely beset with gi-anules, spines or stumpy tubercles ; oral shields usually joined with the first lateral arm plates ; oral angles free of granules. i — Arms strongly knotted, with very prominent spine ridges ; arm spines of both sides of free basal joints approximating dorsally; dorsal arm plates very small, widely separated from each other by the lateral arm plates OpMomit relia. il- — Arms not very knotted ; arm spines of both sides not approximating dorsally ; dorsal arm plates not very small, in contact with each other at least in proximal arm joints. / — Basal tentacle pores not especially large ; single flat tentacle scale to each pore ; radial shields oval OpMopJdhalmus. jj — Several first tentacle pores especially large, while the outer ones are very small ; no true tentacle scales, Avhicli are represented by a number of lowest arm spines much smaller than the upper ones. . . Opldentrema. gg — Radial shields long, narrow, bar-like. k — Radial shields separated from each other; disk, as well as radial shields, more or less closely covered with granules, spines or stumpy tubercles Opliiacantha. JcJc — Radial shields perfectly joined in pairs, naked ; disk sparsely beset with spines or granules, the disk scales being quite distinct Ophiacanthella. ß'- — Radial shields large, long and wide ; disk scales quite distinct. I — Oral j)apilla3 arranged in a regular series. m — Outermost oral papillœ large and operculiform ; arms strongly knotted, with veiy prominent spine ridges ; arm spines of both sides of free basal joints approximating dorsally ; dorsal arm plates very small, widely separated from each other by the lateral arm plates. -n — Disk distinctly five-lobed ; radial shields divergent, those of a pair slightly in contact in the outer parts ; oral shields small, separated from the first lateral arm plates by the adorai shields ; interbrachial ventral surfaces covered with numerous fine scales ; genital bursaa MONOGRArH OF JAPANESE OPHIUROIDEA. 99 rudimentary, being represented merely by the creases between the interbrachial ventral surfaces and arm bases. . , Ophiothamnns . nn — Disk not distinctly five-lobed ; radial shields joined in pairs nearly along the entire length ; oral shields large, joined with the first lateral arm plates ; each interbrachial ventral surface covered chiefly with two or three very large scales ; genital bursœ well developed . . OpMui'othamnus. mm — Outermost oral papilla not especially large; arras not very knotted; arm spines of both sides never approximating dorsally ; dorsal arm -. plates rather large, in contact with each other at least in proximal arm joints ; disk distinctly five-lobed ; special marginal disk scales well developed OpJdoplinfhaca . 11 — Numerous oral papilla not in a single series, but clustered along the sides and at the top of each jaw, or at least on both sides of the outer end of each oral slit. o — Disk distinctly five-lobed ; special marginal disk scales well developed ; radial shields separated from each other ; single tentacle scale to each pore OpMomilra. 00 — Disk not distinctly five-lobed ; special marginal disk scales absent ; radial shields joined in pairs along the entire length ; usually three tentacle scales forming a tube for the ventrally turned up tentacle OpMocamax. ee — Arm plates and arm spines covered over by a cereous skin. p — Radial shields long, narrow, bar -like Ophidlebes. pp — Radial shields short, rounded or triangular OpJdosemnotes. Ophioniyces spathifer Lyman. OpUoimjces spathifer : Lyman, Bull. Mus. Comp. Zool., YI, Pt. 2, 1879, p. 47, PI. XIV, figs. 386-388 ; Lyman, Piep. Challenger, V, p. 240, PI. XIX, figs. 10-12. YensM Sea ; 565 fathoms (Ly^ian). Ophiologimus hexactls Clark Ophiologlmus liexactis : Clark, Bull. U. 8. Nat. Mus., LXXV, 1911, p. 252, fig. 123. 100 ART. 2. — H. MATSUMOTO : One specimen ; Mera-out-Oisegaké, Sagami Sea ; 350 fathoms. One specimen ; off Osliima ; 90-100 fathoms. Off Suno Saki, Sagami Sea ; 83-158 fathoms (Clakk). Though the radial shields are externally invisible, they have been proved to be present by dissection, and are very small, short, bar-like, widely separated from each other. My specimens also invariably show an indication of schizogony, for they are six- armed, three arms being distinctly smaller than the other three. Ophiophruva liodisca (Clark). Opldophrura liodisca : Clark, Biül. U. S. Nat. Mus., LXXV, 1911, p. 249, fig. 121. Off Omai Saki, Yenslni Sea ; 475-505 fathoms (Claek). Ophiopora wegatrenia (Clark). OpUacantha megatrcma : Clark, Bull. U. S. Nat. Mus., LXXV, 1911 p. 287, fig. 113. Off eastern Japan ; 587-943 fathoms (Clark). As stated by Kœhlee, Ophiopora and Ophiotrema are closely related to each other. All the members of these genera, i.e. Ophiopora hartlelil (Lyman, 1883), 0. paucispina (Lütken & MoKTENSEN, 1899), the present species and Ophiotrema alberti Kœhler, 1894, have visible disk scales, large adorai shields, which separate the oral shields from the first lateral arm plates, narrow oral angles, gaping oral slits, very large tentacle pores and distally notched ventral arm plates. I look upon the present species to be an Ophiopora standing next to Ophiotrema, because the distal oral papilla) are differentiated in size from the inner ones, quite like those of Ophiotrema, while the total absence of tentacle scales is a character of Ophiopora. MONOGRArH OF JAPANESE OPHIUEOIDEA. . 101 Ophiambijc acideatus Lyman. OpJiiamhix aculealus : Lyman, Auniv. Mem, Boston Soc. Nat. Hist., 1880, p. 11, PI. n, figs. 29-31; Lyman, Rep. Challenger, V, 1882, p. 235, PI. XXVn, figs. 10-12; Clark, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 114. Colnett Strait; 1,008 fathoms (Clark). Near Fiji Is.; 1,350 fathoms (Ly^ian). Ophiolimna Verrill, 1899, emend. Disk covered with fine granules, the radial shields being usually concealed. Oral shields large, wide, separated from the first lateral arm plates by the adorai shields. Oral angles more or less granulated. Four to six oral papillœ, of which the outer- most one is very large and operculiform. Arms slender, horizon- tally flexible, ^''entrai arm plates wider than long, not in contact with each other. Four to seven arm spines, long, glassy. Single large, leaf-hke tentacle scale to each pore. This genus includes Ophiacantlia baircU Lyman, 1883 (referi'ed to Ophiolimna by Vereill, 1899), Ophioconis antarctica Lyman, 1879, Op)hiaccmtha perfida Kœhler, 1904, Ophiolimna operculata Kœhlek, 1907, Ophioconis diaslata Clark, 1911, 0 papiUata C'Lk'SiK, 1911 and Ophiacantha lambda Clark, 1911. OphiochœtaÇ?) mixta Ly'man, 1878 (referred to Ophiolimna by Verrill, 1899) and Ophiolimna liftoraUs Kœhler, 1912, are in my opinion not genuine Ophiolimna, being different from it in the more numerous oral papillœ, of which the outermost one is small and stretches inwards above the next papilla ; in the well developed ventral arm plates, which are fuUy in contact with each other ; and in the presence of two tentacle scales, of which the abradial 102 ART. 2. — H. MA.TSUMOTO : one overlaps the base of the lowest arm spine. (See under Ophiodermcdiclce.) ILty to Japanese species of OpJnolimna, A — Eadial shields entirely covered over, a — Disk granules very fine ; four to five arm spines rather blunt, some- what longer than the corresponding arm joint diastata. aa — Disk granules coarse, mingled with some scattered spines ; seven acute arm spines, longer ones about thrice as long as the corres- ponding arm joint hairdi. AA — Kadial shields partially naked ; disk granules coarse and elongated ; five acute arm spines, of which the uppermost is the longest and about thrice as long as the corresponding arm joint ........ lambda. O^yhiolhnna diastata (Clark). OjoMoconis diastata : Clark, Bull. U. S. Nat. Mus., LXXY, p. 27, fig. 3. Off Shio Misald ; 244-253 fathoms (Claek). Ophiolimna hairdi (Lyman). Opldacantha hairdi : Lyman, Bull. Mus. Comp. Zool., X, 1883, p. 256, PI. V, figs. 70-72; Lütken & Mortensen, Mem. Mus. Comp. Zool., XXIII, 1899, p. 177, PL Yni, figs. 9-13; Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 230. Ophiolimna hairdi : Yerrill, Bull. Labor, Nat. Hist. State Univ. Iowa, Y, No. 7, 1899, p. 41^^; Yerrill, Transact. Connecticut Acad., X, 1899, p. 345. Off Mikawa, Yenshii Sea ; 943 fathoms (Claek). Bering Sea. Alaska. Washington. Gulf of Panama. Caribbean Sea. Eastern Atlantic. 1) This paper 'was not seen by me. MONOGRAPH OF JAPANESE OPHIUllOIDEA. 103 OpJdolinina lambda (Clark). OpMacantha lambda: Clakk, Bull. U. S. Nat. Mus., LXXY, 1911, p. 231, %. 108. Off Hiuga, southern Japan ; 437 fathoms (Claek). Key to Japanese species of Ophiomitrella, A — Radial shields wider than long ; four oral papillae, of which the outermost one, arising from the first ventral arm plate, stretches above the next papilla ; ventral arm plates not in contact with each other ; seven or eight arm spines, of which the uppermost is the longest stellifera. AA — Radial shields about as wide as long ; six oral papillas on either side ; ventral arm plates slightly in contact with each other ; nine or ten arm spines, of which the uppermost and lowest are the shortest , , , poh/acantjia. Oxthioniitvella stelliferaf sp. nov. Two specimens ; off Inatori, Izu. One specimen ; Locality unknown. Diameter of disk 4 mm. Length of arms 20 mm. Width of arms at base 0.8 mm. Disk almost circular but with a small indentation in each interbrachial border, flat, covered with coarse, stout, partly imbricated scales, which almost invariably bear each a large stumpy tubercle with a six-rayed stelliform crown. Radial shields very small, irregularly triangular, with much rounded inner angle, a little larger than the disk scales, wider than long, sepa- rated from each other. Tnterbrachial spaces below covered with a few coarse, irregular, imbricated scales, without any stumpy tubercles. Genital slits short. 104 ART. 2. — H. MATSÜMOTO. Oral shields large and stout, pentagonal, with acute inner and lateral angles, rounded outer angles, concave outer lateral sides and curved outer side. Adorai shields large, long, stout, three-sided, with round- ed angles and nearly straight adradial and convex abradial sides, meeting each other. Three oral papillae on either side ; the outer- most one is very large, flattened and blunt ; the rest are conical, some- what flattened, obtusely pointed. There is an ad- ditional pointed papilla, which arises from the first ventral arm plate and is directed inwards. Teeth stout, triangular, obtuse. Arms strongly knotted. Dorsal arm plates in the proximal part very small, tlu'ee- sided, with strongly convex outer side and very wide inner angle, much wider than long, with convex surface ; more distally, they become more or less regularly triangular and as long as, or longer than, wide. Lateral arm plates very well developed, strongly constricted at the middle of each joint, those of the two sides meeting: each other both above and below. First Fig. 27. OphiomUrella stellifera. xl2. a. From above. b. From below, c. Side view of two arm joints near disk. d. Side view of two disk tubercles. MONOGRAPH OF JAPANESE OPHIUBOIDEA. 105 ventral arm plate small, pentagonal, with wide and truncated inner angle and curved outer side, wider than long. The rest are large, pentagonal, with wide inner angle and perfectly rounded outer angles, much wider than long. Arm spines long, slender, hyaline, flattened and thorny, except the uppermost one or two which are nearly terete ; upper ones longer ; the uppermost spine is about two to three times, and the lowest one about two-thirds, as long as the corresponding arm joint. From the seventh or eighth joint outwards the lowest spine is transformed into a compound hook, being shghtly bent inwards and bearing series of booklets on the inner side. The arm spines are seven or eight in number and approximated dorsally in the basal joints, but rapidly decrease in number outwards, so that they are four or five in the tenth joint. One tentacle scale, small, oval and flat. Colour in alcohol white. In one of the specimens, the tubercles of the disk are fewer, but in the second, they are larger, and the stelliform crowns are more complex, some of the rays bearing two or three denticles each. In the third specimen, the disk scales are coarser than in the other two, and the radial shields are slightly longer than wide. Tliis new species is near to OpJdomitrella partita (Lütken & MoKTENSEN, 1899), but differs from it chiefly in the smaller radial shields, in the disk tubercles having more slender stem and larger crown, in the arm spines being approximated dorsally in the first basal joints, and in the lowest arm spine being transformed into a compound hook. The six rays of the crown of the disk tubercles are formed by the bifurcation of the three primary rays, so that the disk tubercles are fundamentally of the same plan in both species. 106 ART. 2.— H. MATSÜMOTO : OpJiioniitrella polyacantlia (Claek). OpMomitra pohjacantha : Clark, Bull. U.S. Nat. Mns., LXXV, 1911, p. 187, fig. 86. Eastern Sea ; 103 fathoms (Clakk). Ophiophthalmus, g. uov. Disk covered with more or less imbricating, irregular scales, and may be beset with coarse granules. Eadial shields naked, rounded or oval. Three to six oral papillœ on either side. Teeth arranged in a single vertical row. Dental papillae absent. Dorsal arm plates in contact with each other in the basal arm joints. Numerous arm spines, long, conical, opaque, hardly serrate, never approximating dorsalty. Usually a single large, flat, leaf-like tentacle scale to each pore. This new genus includes Ophiacantha normani Lyman, 1879 (referred to OpUiomitra by Lyman, 1882), OpMomitra granifera LüTKEN & MoETENSEN, 1899, Opliiacanthci rellcta Kœhlee, 1904, Ophlomitrella miitata Kœhler, 1904, 0. languida Kœhlee, 1904, 0. 2olacida Kœhlee, 1904, OpJiiomitra microphyJax Claek, 1911, 0. codonomorpha Claek, 1911, Ophiacantha leucorhahdota Claek, 1911, 0. eurypoma Claek, 1911, 0. hylacanlha Claek, 1911, Ophiomitrella amcricana Kœhlee, 1914, &c., besides the genotype, Ophiacantha cataleimmoida Claek, 1911. The present genus is very nearly allied to Ophiomitrella, but differs from it in the comparatively well developed dorsal arm plates, the basal ones being in contact with each other, and in the arm spines not approximating dorsally in the basal arm joints. Further, this genus differs from the genuine Ophiacantha in the short and rounded radial sliields and in the coarser disk scales. MONOGRAPH OF JAPANESE OPHIUEOIDEA. 107 In my opinion, the shape of the radial shields is an important character in the systematic of the present family. The radial shields of the genuine Opldacantha are long, narrow and bar-like. In Yeerill's key, the species of OpMophtlialmus would appear as belonging to Ophialcœa, but the latter is a close ally of Ophiaccwtlia, with long and bar-like radial shields. Key to Japanese species of OpMo2:)Mhahniis. A — Disk beset with some granules, or sometimes with spines ; a single large tentacle scale to each pore. a — Disk with granules only, without spines. h — Disk granules abundant, uniformly distributed on both the dorsal and ventral sides, c— Kadial shields small, widely separated from each other; several granules occur at the outer border of the dorsal arm plates of basal free arm joints ; fom* or five oral papillae, of which tlie outermost one arises from the first ventral arm plate. d — Disk scales coarse ; outer second oral papilla very large, "s^'ide ; six or seven arm spines cataUimmoidus. dd — Disk scales fine ; innermost oral papilla largest ; four arm spines . . normani. cc — Radial shields large, lying closely side by side in pairs, but not in contact with each other ; six oral papillae on either side ; eight or nine arm spines . . , , leucorhahdotus. bh — Disk granules few, being present only in the central parts of the dorsal surface, radial shields large, widely separated from each other ; three oral papillae on either side ; eight arm spines , codonomorpha. aa — Disk beset Avitli stout rough spines, besides granules ; radial shields moderately large ; four oral papillae on either side ; eight arm spines liylacantlia. AA — Disk beset with many stumpy tubercles with stelhform crowns ; radial 108 ART. 2. — H. MATSUMOTO : shields very small, widely separated from eacli other; six or seven oral papillae on either side, the outermost two arising from the adorai shield; oral shields separated from the first lateral arm plates; six arm spines; a single very small tentacle scale, .microphylax. OpJiiophthalnius cataleimmoidus (Clark). OpUacantlia cataleimmoida : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 217, fig. 100. Three specimens ; off Misaki, Sagami Sea. Numerous specimens, probably the young of the present species ; off Misaki, Sagami Sea. OflrOmaiZaki;475- Jih. 624 fathoms (Clark). Uraga Channel ; 302 fathoms (Claek). Yen- shû Sea, 507 fathoms (Claek). Off Kurile Is.; 229 fathoms (Claek). Bering Sea. Alas- ka. The disk scales are very coarse, stout, par- tially imbricated, or rather tessellated and with distinct sutures, each scale bearing several granules, which are never situated on the sutures, except in the peripheral zones of the Fig. 28. Ophiophthalnius cataleimmoidus, x 5|. a. From above, b. From below, c. Side view of three arm joints near disk. disk, where the sutures are more or less obscured. MONOGRAPH OF JAPANESE OPHIUROIDEA. 109 OphiopJifhahmis norniani (Lyman). OpJiiacanfha normani : Lyman, Bull. Mus. Comp. Zool., VI, Pt. 2, 1879, p. 58, PI. XVI, figs. 414-416 ; Lütken & Moetensen, Mem. Mus. Comp. Zool., XXm, 1899, p. 170, PI. XVI, figs. 1-4 ; Verrill, Trans. Connecticut Acad., X, 1899, p. 349 & 353 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 215. OpJdomifra normani : Lyiman, Rep. Challenger, V, 1882, p. 208, PI. XXVI, figs. 9-11. Eastern Sea; 95-106 fathoms (Claek). Yenshû Sea ; 565-662 fathoms (Lyman, Clakk). Sagami Sea ; 345-775 fatlioms (Lyman, Claek). Off eastern Japan ; 440-629 fathoms (Clakk). South of Hokkaido ; 464 fathoms (Clark). Yezo Strait ; 533 fathoms (Claek). Okhotsk Sea; 75-510 fathoms (Claek). Bering Sea. Alaska. Off Washington. West coast of Mexico. Ophiophthahnus leucorhahdotus (Clark). OpMacantlm, levcorJiabdota : Clark, Bull. Ü. S. Nat. Mus,, LXXV, 1911, p. 222, fig. 102. Eastern Sea ; 103-152 fathoms (Claek). Ophiophthalmus codonomorpha (Clark). Ophiomitra codonomorplia : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 189, fig. 87. Off Mikawa,^ Yenshû Sea ; 943 fathoms (Claek). Ophiophthahnus hylacantha (Clark). OphiacantJia hylacantha : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 227, fig. IOC. Off Omai Zaki ; 918 fathoms (Claek). 110 ART, 2. — H, MATSÜMOTO : Ophiophthahnus niicrohylax (Clark). Opldomitra microliylax : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 184, fig. 84. Eastern Sea ; 103-152 fathoms (Clark). Ophientrema Verrill, 1899. Clark^^ says that his Opliiacantha leucostida " is almost certainly entitled to be the type of a new genus, the large amount of uncalcified skin and the peculiar condition of arm spines and tentacle scales furnishing good generic characters." As far as I can judge, the species in question belongs to Veeeill's subgenus Ophientrema, being identical with Ophiacantha [Ophientrema) scolopendrica Lyman, 1883. Though Ophientrema was established as a subgenus of Ophiacantha, I am inclined to rank it as a distinct genus, because the shape of the radial shields, the large basal tentacle pores, the peculiar lower arm spines and tentacle scales, &e. are different from those of the genuine Ophiacantha. This procedure would naturally satisfy the need felt by Claek. Key to species of Ophientrema. ^ — Disk covered with fine imbricating scales, and with minute, thorny, scattered tubercles euphylacteiim. ^A. — Disk covered by a soft skin, scales being distinct only near the radial shields, and Avith minute, scattered granules . . scolopendricum. Ophientreina eiipliylacteutn, Ophiacantha euphjladea : Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 225, fig. 105. 1) Bull. U. S. Nat. Mus., LXXV, ISHl, p. 230. MONOGRAPH OF JAPANESE OPHIUROIDEA. Ill Off Manazurn Zaki, Sagami Sea ; 120-265 fathoms (Clark). Ophientrema scolojyendviciini (Lyman). Opidacantha scolopendrica : Lyman, Bull. Mus. Comp. Zool., X, IS 83, p. 259, PI. VI, figs. 85-87. OpMacanllia ( Ophientrema) scolopendrica : Verrill, Transact. Connecticut Acad., X, 1899. p. 332. Opiiiaccmtha Jeucosticta : Clark, Bull, U. S. Nat. Mus , LXXY, 1911, p. 235, fig. 111. Yensliû Sea ; 565 fathoms (Lyman), 507 fathoms (Claek). Key to Japanese species of OphiacantJia. A — Oral papillse arranged in a single series, a — Single tentacle scale to each x^oi'e» b — Arm spines smooth. c — Disk covered with spines or stumpy tubercles. d — Disk tubercles ending with a thorny crown, e — Four to seven arm spines. /—Four arm spines, which are shorter than the corresponding arm joint dallasii. ff — ^Five or more arm spines, of which the u^^per ones are distinctly longer than the corresponding arm joint. g — Five or six arm spines; outer end of the radial shields ■ns^eà..(Bnigmatica. gg — Seven arm spines ; radial shields entirely covered over. h — Outermost oral papilla very large and wide. i — Outermost oral papilla much wider than long ; tentacle scales acute . . levisplna. ii — Outermost oral papilla longer than wide ; tentacle scales large, flat, blunt omoplata. M — Outermost oral papilla not especially large and wide. ' ./ — Oral shields wide, wdth acute lateral angles ; tentacle scales very ^,^ small, spiniform and acute pentagona. 112 ART. 2. — H. MATSUMOTO : jj — Oral shields not very wide, with rounded lateral angles ; tentacle scales oval, flat, bluntly pointed adiapJiora. ee — Ten or eleven arm spines ; stumpy tubercles of the disk very, short ; oral shields wide ; tentacle scales large, long, acute ancMlahra. dd — Disk covered only with spines. h — Three oral papillae on either side ; no spines along the outer border of basal dorsal arm plates. I — Disk spines smooth ; outermost oral papilla very long and spiniform ; eight or nine arm spines ; tentacle scales conical, acute lophohracliia. II — Disk spines rough ; outermost oral paj)illa flat and clavate ; ten or eleven arm spines ; tentacle scales long and flat acanthinotaia. kk — Six oral papillae on either side ; a row of spines present along the outer border of basal dorsal arm plates ; eight arm spines ; tentacle scales acute inutilis. cc — Disk covered with granules ; seven or eight arm sj)ines ; tentacle scales large hiderdata. bb — Arm spines serrate. m — Oral papillae and tentacle scales smooth. n — Disk covered with fine scales and sparsely beset with rough, stout spines ; six arm spines .prionota. nn — Disk covered with thorny, stumpy tubercles, among wliich several long, thorny spines are mingled ; nine arm spines diploa. mm — Oral papilla and tentacle scales rough; disk tubercles ending with a thorny crown ; six to eight arm spines o rhachôpîiora. act —Two tentacle scales to each pore ; disk covered with fine granules ; six arm spines, flattened, smooth hisquamata. A A — Oral papillae not arranged in a single series, but the outer ones clustered around the second oral tentacle pore ; disk co%'ered with short, granule-like spines ; twelve serrate arm spines ; single tentacle scale to each pore rosea. MONOGRAPH OF JAPANESE OPHIUROIDEA. 113 Ophiacantha dallasH Duncan. OpJdacantha dullas'ii : Duncan, Journ. Linu. Soc. London, XIY, 1879, p. 471, PL XI, figs. 2Ö-27 ; Lyman, Rep. Challenger, V, 1882, p. 199 ; DÖDERLEIN, Semon-ZooI. Forscliungsr. Australien u. Malayisclien Archipel, V, 1896, p. 291, PL XrV, figs. 3a-3c, PL XVI, figs. 12-12a. Off Korea, Sea of Japan ; 50 fathoms (Duncan). Amboina (Dodeelein). Ophiacantha œnigmaticaf sp. uov. Seven specimens ; oft' Misaki, Sagami Sea. Diameter of disk 5 mm. Length of arms 25 mm. Width of arms at base 1.2 mm. Disk penta- gonal, with con- cave interbrachi- al borders, flat, thin, covered by a thick skin, con- taining rather fine, rounded, thin, irregular scales, which be- come distinct when dried ; sparsely beset with small stumpy tuber- cles, WniCn are pjg_ 29. Ophiacanth'i œiùijinatlca. xl2. a. From aljove. b bifid nv Pvnwn- From below, e. Side view of two arm joints near disk. 114 ART. 2. — H. MATSUMOTO: ed with a few thorns. Eadial shields raised, long, narrow, about two-thirds as long as the disk radins, those of the same pair parallel to each other. They are covered by the skin, and usually also by the scales except at the outer ends. In some specimens, they are almost entirely fi^ee from the scales. Interbrachial ventral surfaces similar to the dorsal side, but the stumpy tubercles are fewer and finer. Genital slits large, long, almost reaching the disk margin, much widened and rounded at the inner ends. Oral shields small, rhomboidal, with concave sides and perfectly rounded lateral and outer angles, inner angle like a beak, so that the inner sides form a brace-shape ; very short, about twice as wide as long. Adorai shields large, stout, with concave adradial side, more or less separated fi'om each other by a naked, depressed space, or sometimes entirely in contact. Three or four oral papillœ on either side, long, conical, rather obtuse ; the outermost one is directed laterally and protects the second oral tentacle pore ; the next is the stoutest. Five or six teeth, obtuse or sometimes widened at the free end, with two or three irregular denticles, except the uppermost one, which is acute and is the longest. Arms rather stout, uniformly tapered, knotted, with very long joints. First two dorsal arm plates quadrangular, with very convex outer side, wider than long, in contact with each other. The second is the largest, and is wider without than witliin. Those following are small, triangular, with very convex outer side, slightly longer than wide, not in contact with each other. Lateral arm plates well developed, with prominent spine ridges, meeting each other aljove and below, except between the first and second ventral arm plates and between the first and second dorsal plates ; besides, the second and third dorsal plates often do not meet, MONOGEAPH OF J.\PANESE OPHIUEOIDEA. 115 leaving a long, narrow, naked space. First ventral arm plate small, hexagonal, with concave inner side, about as wide as long. The second is the largest, and is triangular, with truncated inner angle and curved outer border, a little wider than long, in contact with the first plate. The rest are triangular, with acute inner angle and curved outer border, wider than long, separated from each other by the lateral arm plates. The outer border of the ventral plates consists indistinctly of three portions, of which the median is very slightly concave. Five or six arm spines, long, slender, spiniform, terete, solid, entire, except the lower spines, which are slightly thorny at the tips. The upper spines are longer than the lower. The uppermost one is about twice as long as, and the lowest a little shorter than, the corresponding arm joint. The lower spines are crowded and lie flat on the ventral surface of the arms. One tentacle scale, very minute» acutely pointed. Colour in alcohol : disk light yellowish gray, arms light yellow. This species is very near to 0. frachijbaclra Claek, 1911, but differs from it cliiefly in the fewer disk tubercles, in the fewer oral papillae, in the shape of the oral shields and ventral arm plates, and in the fewer arm spines, which are never club-shaped. Like 0. trachybadra, this species has some external re- semblances to Ophiolehes. However, I believe with Claek, that it is more natural to plat3e it in Ophlacantha than in Ophiolehes. Ophiacantha levispina Lyman. Opiiiacantha levispina : Lyman, Bull. Mus. Comp. Zool., V, 1878, p. 147, PI. X, fig. 277 ; Lyman, Rep. Challenger, V, 1882, p. 196, PI. XXV, figs. 1-3; Clark, BuU. U. S. Nat. Mus., LXXV, 1911, p. 198. 116 ART. 2. H. MATSUMOTO ! Sea of Japan ; 224-248 fathoms (Clark). Off Kii, Kumano Sea; 507-649 fathoms (Clark). Off Omai Zaki ; 624 fathoms (Clark). OffHiuga; 720 fathoms. Eastern Sea ; 361-440 fathoms (Clark). Malaysian waters. Ophiacantha onioplata CYark. Ophiacantha omoplata : Clark, Biül. U. S. Nat. Mus., LXXV, 1911, p. 202, iig. 93. Off Korea ; 136 fathoms (Clark). Ophiacantha pentagona Kœhler. OpMacantha pentagona: Kœhler, Ann. Sei. Nat. ZooL, 8" Ser., IV, 1897, p. 342; Kœhler, Ech. Indian Mus., Deep-sea Oph., 1899, p. 53, PI. IV, figs. 27-29; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 196. OpMacantha pentagona var. armata : Kœhler, Kés. Camp. Sei. Monaco, Xn. 1198, p. 55. Numerous specimens ; Uraga Channel. Numerous specimens ; Sengenzuka-Aoyamadashi, Sagami Sea ; 85 fathoms. Numerous specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Numerous specimens; off Oshima ; 90-100 fathoms. Off lui, Kumano Sea ; 191-253 fathoms (Clark). Off Mikawa, Yenshû Sea; 943 fathoms (Clark). Off Port Heda ; 161-167 fathoms (Clark). Sagami Sea; 52-153 fathoms (Clark). Suruga Gulf; 45-131 fathoms (Clark). Off Omai Zaki; 918 fathoms (Clark). Eastern Sea ; 95-369 fathoms (Clark). Korea Strait ; 59 fathoms (Clark). Malaysian waters. West of Africa. MONOGRAPH OF JAPANESE OPHIUIIOIDEA. 117 Ophiacantha adiaphora Clark. OjpMacantha adiaphora: Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 199, fig. 91. Sagami Sea ; 622 fathoms (Clark). Sea of Japan ; 90-207 fathoms (Clark). Okhotsk Sea ; 68 fathoms (Clark). Bering Sea. Ophiacantha anchilabra Clark. Ophiacantha anchilabra: Clark, Bull. U.S.Nat. Mus., LXXV, 1911, -jx 205, fig. 95. One specimen ; locality unknown, probably Sagami Sea. Off Kii, Kumano Sea ; 649 fathoms (Clark). Off Omai Zaki ; 918 fathoms (Clark). Ophiacantha lophohrachia Clark. Ophiacantha lophohrachia : Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 232, fig. 109. Eastern Sea; 152 fathoms (Clark). Ophiacantha acanthinotata Clark. Ophiacantlux acanthinotata : C*lark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 203, fig. 94. Numerous specimens ; Okinosé, Sagami Sea. Two specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Gulf of Tokyo; 169 fathoms (Clark). Eastern Sea; 181 fathoms (Clark). Ophiacantha iniitilis Kœhler. Ophiacantha inutilis : Kœhler, Exp. Siboga, XLV, PI. 1, 1904, p. Ill, PI. XXI, figs. 6-8 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 208. 118 ' AKT. 2. — H. MATSÜMOTO : Off Kii, Knmano Sea; 191 fathoms (Clark). Uraga Channel ; 70-197 fathoms (Claek). Eastern Sea ; 440 fathoms (Clark). Malaysian waters. O^yhiacantJta hldentata (Retzius). Asterias hidentata : Retzius, Diss., 1805, p. 33.^^ Opldura retzii : NiLSSON, Coll. Zool. Scand., 1817, j). 15.^^ OpUocoma ardica : Müller & Troschel, Sys. Ast., 1842, p. 103. Opliiacantha splmâosa : Müller & Troschel, ibid., p. 106 ; Lütken, Addit. ad Hist., I, 1861, p. 6-5, PI. 11, fig. 14 ; Lyman, 111. Cat. Mus. Comp. Zool, I, 1865, p. 93, figs. 6 & 7. OpJdacaniha grönJandica : Müller k Troschel, Arch. Natm-g, 1844. p. 183.^> OpJiiocoma ecldnulaia : Forbes, Sutherlind's Journ. Voy. Baffin's Bay n, App., 1852, p. 205.' > OpUacantJia hidentata : Ljungman, Öfv. K. Vet. Akad, Forh. XXVIII, 1871, p. 652'^; Lyman, Bull. Mus. Comp. Zool., Y, 1878, p. 149; Kœhler, Res, Camp. Sei. Monaco, XII, 1898, p. 55 ; Grieg, Fauna Arctica, I, 1900, p. 267; Rankin, Proc. Acad. Nat. Sei. Philadelphia, 1901, p. 179; Norman, Ann. Mag. Nat. Hist., 7th. Ser., XH, 1903, p. 407 ; Kœhler, Mem. Soc. Zool. Fr., XIX, 1906, p. 6 ; Kœhler, Exp. Sei. Travaillem' et Talisman, yJH, 1907, p. 289; Kœhler, Bull. Sei. Fr. Belg., XXXIV, 1907, p. 317; Kœhler, Bull. Mus. d'Hist. Nat., 1909, x>. 123 ; Kœhler, Res. Camp. Sei. Monaco, XXXIV, 1909, p. 184; Kœhler, Bull. Mus. d'Hist. Nat., 1909, p. 123 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 195 ; Kœhler, BuU. Mus. d'Hist. Nat., 1913, p. 14 ; Kœhler, Bull. U. S. Nat. Mus., LXXXIV, 1914, p. 80. One specimen ; locality unknown. Sea of Japan ; 130-536 fathoms (Clark). Grulf of Tartary 318 fathoms (Clark). Off Knrile Is.; 229 fathoms (Cl.\rk). 1) These papers were not seen by me. MONOGEAPH OF JAPAî^ESE OPHIÜEOroEA. 119 Bering Sea. Arctic Ocean. North Atlantic. Opliiaeantha prionota Claek. Opldacantha prionota : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 213, fig. 99. Off Tanega-shima ; 1,008 fathoms (Clark). Ophiacantha diploa Claek. OpJdaeantha diploa: Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 207. Off Hiuga ; 437 fathoms (Clark). OpMncantha rhachophora Clark. Ophiacantha rhachophora : Claek, Bull. U. 8. Nat. Mus., LXXV, 1911, p. 201, %. 92. Four specimens ; off Inatori, Izn. Eastern Sea ; 106-139 fathoms (Clark). Off Omai Zald ; 475- 505 fathoms (Clark). Off Ose Zaki, Sm^nga Gulf ; 63-100 fathoms (Clark). Sagami Sea ; 83-153 fathoms (Clark). Uraga Channel ; 197 fathoms (Clark). Off Kii ; 191-253 fathoms (Clark). Sea of Japan ; 90-207 fathoms (Clark). Bering Sea. I hesitated at first to refer these specimens to the present species. The distal end of the radial shields is naked but not conspicuous, and the outermost oral papilla arises from the adorai shield. Besides, there occurs one scale-like papilla on the first ventral arm plate, projecting inwards and vertically. These characters are not apparent in Clark's figure ; but it is to be remarked that the insertion of the outermost oral papilla is very liable to be misunderstood when the papilla is not turned up but 120 AET. 2. — H. MATSUMOTO Fig. SO. Ophiacantha rhachophora. x 18. a. From above, h. From below. 0. Side view of two arm joints near disk. directed in the same direction with the others, so that in such a case the oral papillae appear as shown in Claek's figure. OpJiiacantJia bisquamata Matsumoto. Ophiacantha hisquamata : Matsumoto, Proc. Acad. Nat. Sei. Plüladel- phia LXVn, 1915, p. 62. MONOGRAPH OF JAPANESE OPHIUEOIDEA. 121 Two specimens ; off Oshima, Sagami Sea ; 75-85 fathoms. Diameter of disk 6 mm. Length of arras 34 ram. Width of arms at base 1.5 mm. Disk pentagonal, with nearly straight or shghtly convex interbrachial borders, flat, soft, closely covered with fine granules, of which eight or nine are contained in 1 mm. Radial shields entirely covered over, very small, bar-like, separated from eacli other. Interbrachial ventral sm-faces similar to the dorsal side, but fi'ce of grannies and covered with fine scales in the inner parts. Genital slits long, nearly reaching the disk margin. Oral shields small, rhom- boidal, with convex inner sides and rounded outer angle, nearly as long as, or slightly longer than, wide, joined with the first lateral arm plates. Adorai shields small, triangular, pointed inwards, meeting with each other. Five or six oral pa- pilla? on either side of each jaw ; the outermost two are flat, leaf-like and protect the second oral tentacle pore, while the others are very narrow and acute ; the innermost one, which forms a pair with that of the other side, is infradental. Four or five teeth in a single vertical row, more or less stout, obtuse. Fig. 31. Ophiacantha hisquamata. x8. a. Fi-om above, h. From below. c. Arm spines of one side of an arm joint near disk. 122 ART. 2. — H. MATSUMOTO : Arms composed of short and wide joints, uniformly tapered. Dorsal arm plates rhomboidal, with very obtuse outer angle, widei- than long, slightly in contact with each other, with a more or less distinct median keel, so that the dorsal side of the arm is keeled as a whole. Lateral arm plates Avith prominent spine ridges, meeting neither above nor below. First ventral arm plate very small, quadrangular, with concave inner side, longer than wide. Those following are moderately large, pentagonal, witli convex but slightly notched outer side and rounded outer angles, about as wide as long. Six arm spines, long, flattened, more or less curved, truncate, translucent, but not serrate ; the uppermost or the next spine is the longest and is about two and a half times as long as the corresponding arm joint, while the lowest one, the shortest, is slightly longer than the same. Two oval, thin, leaf -like tentacle scales to each pore. Colour in alcohol : disk grayish brown ; one specimen with five white patches, which correspond to the radial shields, the other without any ; arms banded with grayish brown and white. Tlie grayish brown parts are dark green and the white parts vivid red in life, according to my own observation on board the " Ikuomara." This species more or less resembles 0. bidentata in the disk coverings, though precise comparison may be dispensed with. The presence of two tentacle scales to each pore, together with the very fine granulation, makes the present species very easy to be recognised. Ophiacantha rosea Lyman. Opldacantha rosea : Lym.\n, Bull. Mus. Comp. Zool., V, 1878, p. 139, PL X, figs. 267 & 268; Lyman, Rep. Challenger, V, 1882, p. 148, PL XXV, figs. 10-12. MONOGRAPH OV JAPANESE OPHIÜEOIDEA. 123 Sagami Sea; 420-775 fathoms (Lyman). ! Lat. 46° 40' S., long. 37° 50' E.; 310 fathoms (Lit^ian). ! Lat. 50° 10' S., long. 74° 42' E,; 175 fathoms (Lyman). Opliiacantltella acontoiyhora (Clark). Opldomitj-a acontopJiora : Clark, Biül. U. S. Nat. Miis., LXXV, 1911, p. 190, %. 88. Off Kiirile Is.; 229 fathoms (Claek). Bering Sea. Alaska. Off Aleutian Is.; 1,213 fatlioms (!) (Claek). This species appears to be distinguished fi'om the genotype, 0. troscheli (Lyman, 1878), by the total absence of the disk granules, by the more numerous disk spines, which have a few denticles near the tip, by the fewer oral papillae and arm spines, and by the smaller dorsal arm plates, which are not in contact with each other. I consider that, it is more natural to refer this species to the present genus than to any other known one. Revision of Ophiothamnus, s. ext. It is a wonder that, the present genus has been long left in a great confusion. As far as I can judge, those species which have been referred to the present genus by modern systematists correspond to a type, which was not considered to be Ophiothamnus by Lyman, while certain species, which must be congeneric with Lyman's type of Ophiothamnus, have been referred to various other genera. For example, the so-called " Ophiothamnus " of modern systematists includes Opkiomitra exigua Lyman, 1879 (referred to Ophiothamnus by Verkill, 1899), Ophiothamnus lœvis Lütken & MoETENSEN, 1899, 0. stultus Kœhlee, 1904, &c., while the genuine Ophiothamnus, referred to other genera by modern systematists, 124 ART. 2. — H. MATSUMOTO : includes Opliîoleda minima Kœhlee, 190ß {=^ Ophioplinthaca occlusa Kœhlee, 1907), Ophiomitra hàbrotata Clark, 1911, &c. The differences between these two groups may be given as follows : Genuine Ophiothamniis : — Disk marked out into five brachial lobes by deep interradial notches of the interbrachial borders, and strongly puffed out interbrachially ; disk scales fine ; radial shields not perfectly joined, but in contact with ■ each • other only at the outermost parts (except in 0. remotus)', oral shields very small, separated from the first lateral arm plates by the adorai shields, which are very large ; interbrachial ventral surfaces large, covered with numerous fine scales ; arms very slender, composed of long and slender joints ; arm spines not serrate, the lowest one being never hook-shaped ; ventral arm plates very small, much narrower than the corresponding arm joint ; genital bursas very rudimentary, being represented merely by the creases . between the interbrachial ventral surfaces and the arm bases ; genital plates very closely set in radial pairs, lying above the basal vertebrœ ; genital scales absent ; generative glands covered by a membrane, which contains fine scales ; peristomial plates triple ; distal vertebrae incompletely divided into halves by a longitudinal series of pores. So-called Ophiothamnus : — Disk not distinctly lobed ; disk scales coarse ; radial shields joined in pairs along nearly the entire length ; oral shields large, in contact with the first lateral arm plates ; adorai shields large, lying entirely proximal to the oral shields ; interbrachial ventral surfaces small, covered with rather few coarse scales, of wliich two or three are large and conspicuous ; arms stout at the base, composed of short and stout joints ; arm spines usually MONOGRAPH OF JAPANESE OPHIUROIDEA. 125 serrate, the lowest one being hook-shaped in distal arm joints ; ventral arm plates large, wide, nearly as wide as the corresponding arm joint ; genital bursae well developed ; genital plates, genital scales, generative glands and peristomial plates unknown, but probably normal ; distal vertebrae unknown. In view of these differences, the two groups can hardly be united in a single genus. I therefore propose the name Ophiuro- thamnus, nov., for the so-called Ophiofhamnus, to distinguish it fi'om the genuine Ophiotfiamnus. There are two other species, viz. Ophiothamnus gracilis Studer, (1882) 1883, and Ophiacantha gracilis Vereill, 1885 (referred to Ophiothamnus by V., 1899), which have been referred to Ophiothamnus, but probably belong neither to the genuine Ophiothamnus nor to Ophiurothamnus. Though I am not able to settle their systema- tic position owing to their very imperfect descriptions, they may possibly be referable to Ophiacantha. Ophiothamnus Lyman, 1869, emend. Sjn. OpMoleda Kœhler, 1906. Disk divided into five brachial lobes, covered with fine, imbricating scales, and beset with scattered spines. Eadial shields large, naked, more or less joined in pairs outwards. Genital bursœ very rudimentary, being represented by the creases between the disk and the arm bases. Oral shields small, separated by the large adorai shields from the first lateral arm plates. Three or four oral papilla3 on either side of each jaw, the outermost one being very large and operculiform. Peristomial plates very large, triple. Arms inserted ventrally to the disk, composed of long, slender, hour-glass-shaped joints. Arm spines long, slender, spini- 126 ART. 2. — H. MATSUMOTO : form, acute, glassy, but not serrate, the lowest one being never hook-shaped. This genus includes Ophiothammis ricariiis Lyman, 1869, 0. qßnis Ljungman, 1871, 0. remotus Lyman, 1878, Ophioleda minima Kœhlee, 1906, Ophiomitra habrotata Claek, 1911 and Ophiothamnus venustus Matsumoto, 1915, the first being the genotype. I consider Ophioleda Kœhlee as a synonym of Ophiothamnus Lyman, because I am not able to make any essential distinctions of generic character between the two genotypes, judging from the descriptions and figures of Lyivian and Kœhlee. Key to Japanese species of OpMotliamniis. A — Disk spines dimorphic, a few very long, acute spines being present, besides very minute ones ; interbracliial ventral svufaces entirely free of spines ; dorsal arm plates comparatively large, longer than half the corresponding arm joint in free Imsal joints ; eight or nine arm spines hahrotatus. AA — Disk spines monomorphic, very minute ; interbrachial ventral surfaces also beset Avith several spines ; dorsal arm plates very small, much shorter than half the corresponding arm joint ; nine or ten arm spines vennstKS. Oiyhiothamnûs hahfotatus (Claek). OpMomiira hahrotnta : Clark, Bull. Ü. S. Nat. Mus., LXXY, 1911, p. 182, fig. 83. Eastern Sea ; 95-lOG fathoms (Claek). OpJiiothainmis venustus Matsumoto, 1915. OpJiiofhamnns venustus : Matsüäk/io, Tree. Acad. Nat. Sei. Philadelphia, 1915, p. 63. Numerous specimens,; off Inatori, Izu, Sagami Sea. MONOGRAPH OF JAPANESE OPRIUROIDEA. 127 Diameter of disk 3 mm. Lengtli of arms 23 mm. Width of arms at base 0.5 mm. Disk five-lobed, i.e. divided by five radiating interradial f arrows into five heart-shaped lobes, with large radial shields and Fig. 32. Ophiothamnus venustus. x20. a. From above, h. From below. c. Side view of three arm joints near disk. 128 ART. 2. — H. MATSÜMOTO : fine, imbricating scales, the latter bearing here and there very short, acnte spines. The radial shields are very large, triangular, witli rather acnte inner and perfectly rounded abradial angles ; the adradial side is the longest, three-fourths as long as the disk radius ; the shields are about twice as long as wide, and are apposed to each other by the outer one-third to one-fourth of the adradial sides. Interbrachial ventral surfaces covered by fine, imbricating scales and beset with short, acute spines ; the latter being less numerous than on the dorsal surface. Oral shields very small, triangular, with acute inner and rounded outer angles, slightly longer than wide. Adorai shields very large, long, stout, wider outwards, tapered inwards, meeting with each other. Oral plates exceedingly small. Three or rarely four oral papillae on either side ; the outermost one arises from the adorai shield and is very large and unusually wide ; the rest are very small ; there is often one additional infradental papilla. Teeth four, blunt, with widened free end, except the uppermost one, which is the longest and is acute. Arms slender and long, strongly knotted. Dorsal arm plates very small, three- sided, with extremely wide inner angle and curved outer border, extremely short, much wider than long. Lateral arm plates extremely well developed, meeting each other above and below, and so well soldered below that the line of suture is hardly visible ; with well developed spine ridges ; strongly constricted at the middle. First ventral arm plate exceedingly small, pentagonal. The rest are quadrangular, with curved outer side, wider than long, decreasing in size outwards. The arm spines are needle-like, very long, acute. They are nine or ten in number and approximated dorsally in the first two or three free joints. In the first free joint, the uppermost spine is about three MONOGRAPH OF JAPANESE OPHIUROIDEA. 129 times, and the lowest one about one and a half times, as long as the corresponding arm joint. More distally, the arm spines become rather rapidly fewer and shorter, so that in the tenth free arm joint, they are five in number and about as long as the correspond- ing arm joint. One tentacle scale, comparatively large, blunt in the basal joints, but pointed more distally. Colour in alcohol white. I have observed the internal structures of this species, and ascertained that the internal oral skeleton, genital plates and generative glands are quite similar to those of the genotype (Lyman, Challenger Kep., PL XLII, fig. 1). Ophiurothamnus, g- nov. Disk covered with rather coarse scales. Radial shields large joined in pairs nearly along the entire length. Interbrachial ventral surfaces small, with a pair of large, distinct plates joined to each other in the interradial line. Another large, prominent plate is present at the middle of the interradial margin, lying just outside the paired plates mentioned. Genital bursse well developed. Oral and adorai shields stout, the latter being entirely proximal to the former. Three or four oral papillae on either side, the outermost one being usually very large and wide. Teeth arranged in a single vertical row. Dental papiUse absent. Arms strongly knotted, with hour-glass -shaped joints. Dorsal arm plates very small. Lateral arm plates well developed, those of the two sides meeting above as weU as below. Ventral arm plates short, very wide, as wide as the corresponding arm joint. Arm spines long, slender, usually serrate ; those of the first one or two free arm joints approximated dorsally. Lowest spine of the outer arm joints hook-shaped. One tentacle scale to each pore. 130 ART. 2. — H. MATSÜMOTO : The present genus includes Ophiomltra exigua Lyman, 1879 (referred to Ophiothamnus by Verrill, 1899), Ophiothamnus Icevis LtJTKEN & MoRTENSEN, 1899, and Ophiothamnus stultus Kœhler, 1904, besides the genotype, Ophiomltra dicycla Clark, 1911. Ophiurothamnus is very near to Ophiomi/tis and Ophtojjliuthaca, but differs from the former chiefly in the very small dorsal arm plates and in the basal arm spines approximating dorsally ; from the latter in the absence of the marginal disk scales, in the smaller interbrachial ventral surfaces, which are covered with fewer, coarser scales, in the smaller dorsal arm plates, and in the basal arm spines approximating dorsally. Further, this genus differs from Ophiothamnus in the coarser disk scales, in the radial shields being joined along the whole length, in the fewer and coarser scales of the interbrachial ventral surface, in the well developed genital bursae, in the stout oral shields, which are in contact with the first lateral arm plates, in the stouter arms, in the very wide ventral arm plates, which are as wide as the corresponding arm joints, in the serrate arm spines, and in the usually hook-shaped lowest spines in the distal arm joints. OphiurotJiamnus dicyclus (Clark). Ophiomltra dicycla : Clark, Bull. U. S. Nat. Mas., LXXV, 1911, p. 181, fig. 82. Eastern Sea ; 434 fathoms (Clark). Bungo Channel ; 437 fathoms (Clark). OpJiiopHntJtaca cardiomorpha (Clark). Ophiomltra cardiomorplia : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 179, fig. 81. MONOGRAPH OP JAPANESE OPHIUROIDEA. 131 Eastern Sea; 3G1-440 fathoms (Clauk). Bungo Channel; 720 fathoms (Clark). Ofl" Shio Misaki ; 587 fathoms (Clabk). Off Omai Zaki ; 624 fathoms (Clark). Key to Japanese species of OpUiomitra, A — Disk scales coarse; radial shields very narrow, widely separated from each other ; oral shields rhomboidal or pentagonal, abont as wide as long ; five or six arm spines, of Avhich the uppermost or the next one is the longest and al>out twice as long as the cor- responding arm joint hytJdaspis. ^^— Disk scales fine ; radial shields not very wide, lying rather closely side by side ; oral shields much wider than long ; six or seven arm spines, of which the uppermost or the next one is the longest and about tlnice as long as the corresponding arm pint , Jifhosora. OpJiioniitrn hythiaspis Clark. OpMomitra hythiaspis : Clark, Bull. U. 8. Nat. Mus., LXXV, 1911, p 185, fig. 58. Off Yenshii Sea ; 043 fathoms (CJlark). OpJiiomitra lithosora (Clark). OpUocamax lithosora : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 191, fig. 89. Off Kii ; 244-253 fathoms (Clark). Eastern Sea ; 361 fathoms (Clark). In the genuine OpUocamax, the disk is not very deeply lobed and the radial shields are joined in pairs along the whole length. The present species lacks these characteristics of Ophlocamax, but exhibits all the characteristics of Ophiomitra as emended by Verrill, 1899. 132 ART. 2. — H. MATSUMOTO : Opliiocamax polyploca CJlark. OpUocamax pdlyploca : Claek, Bull. U. S. Nat. Mus., LXXY, 1911, j). 193, %. 90. Eastern Sea; 95-152 fathoms (C'lakk). Kagoshima Gnlf ; 85 fathoms (Claek). Key to Japanese species of OpJtioiebes, A— Four oral papillge od either side of each jaw ; ventral arm plates not notched outwards at least within the disk ; seven arm spines ; single tentacle scale to each pore ; disk beset with numerous small, rather well spaced, stumpy tubercles hrachygnatha. AA — Three oral papillae on either side ; ventral arm plates distinctly notched outwards even witliin the disk ; five arm spines ; tentacle scales absent. a — Disk closely beset with very numerous small, rather granule-like, stumpy tubercles asaphes. aa — Disk sparsely beset with rather large, conical tubercles, which are more or less numerous on the radial shields, but few in the other parts tnherosa. OpJiiolebes brachygnatha Clark. OpMolehes hrachygnatha : Clark, Bull. U. 8. Nat. Mus., LXXY,^1911, p. 240, fig. 115. Off Simushir Is.; 229 fathoms (Clark). Ophiolehes asaphes Clark. OjoMokhes asaphes: Clark, Bull. U. 8. Nat. Mus., LXXV, 1911, p. 239, fig. 114. OffKinkwasan; 182 fathoms (Claek). Off Hokkaido ; 175- 349 fathoms (Clark). Off Saglialin ; 100 fathoms (Claek). MONOGllAPH OF JAPANESE OPHIUROIDEA. 133 Ophiolehes tiiberosa Matsumoto. OpMolebes tvberosa : Matsumoto, Proc. Acad. Nat. 8ci. Pliilaclelphia, 1915, p. 64. Numerous specimens; Okinosé (a submarine bank), Sagami Sea. Diameter of disk 10 mm. Length of arms 38 mm. Width of arms at base 1.5 mm. Disk five-lobed, with strongly concave interbrachial borders, deeply hollowed at the central region, covered by a thick, cereous skin, wliich contains well spaced, coarse granules or thick, rounded scales of various size ; beset with several short, conical, #^Ä^^.? Fig. 33. Ophiolehes hiberosa. x 8. a. From above, h. From below, c. Side view of three arm joints nenr disk. 134 ART. 2. — H. MATSÜMOTO : stout, obtuse tubercles, which are larger and more numerous on the radial shields. Radial shields also covered by the skin, long, narrow, strongly raised, about two-thirds as long as the disk radius. Inter brachial ventral surfaces covered by a skin similar to that of the dorsal surface, the granules and tubercles being however smaller. Genital slits large, long, but not reaching the disk margin. Oral skeleton also covered by the cereous skin. Oral shields rather small, thick, rhomboidal, wider than long, with wide, rounded outer angle and convex surface. Adorai shields large, quadrangular, with perfectly rounded outer angles and strongly convex surface, wider without than within, meeting wdth each other inwards. Between each pair of oral plates, occurs a more or less distinct buccal pore. Three or sometimes four oral papillae on either side, conical and blunt ; inner ones smaller ; the outer- most papilla is very large and stout. There occurs often one additional papilla just below the teeth. The oral papillae project laterally so as to reach beyond the radial axis, and those of the two sides of each oral slit are placed alternately. Teeth conical, stout, obtuse. Arms slender, covered by a thin, cereous skin. Dorsal arm plates two to each joint ; the inner one has convex surface, and is small, quadrangular, wider than long, but in the more distal part of the arm is often longer than wide ; the outer one is large, fan-shaped, much wider without than within, and has convex surface. The dorsal side of the basal region of the arms is covered by the continuation of the disk covering, so that it bears coarse, rounded, smooth, convex, well spaced granules or scales of various size in place of the dorsal arm plates. Lateral arm plates somewhat flared outwards, meeting below for a short MONOGliArH OF JAPANESE OrHIUROJDEA. 135 extent. First ventral arm plate comparatively large, hexagonal, with concave inner side and convex surface, widest at the lateral angles, as long as or slightly longer than wide, in contact with the next plate, which is the largest of all, pentagonal, widest at the lateral angles, as long as or slightly longer than wide, and has ■ convex surface and a conspicuous notch at the outer angle. The rest separated from one another, rhomboidal, with a con- spicuous notch at the outer angle, with strongly convex surface ; more distally they become smaller, oval or round, and the surface is so convex as to appear like hemispheric^ tul)ercles. The arm spines are five in number in the proximal joints, but four in the more distal ones. They are conical, blunt, solid, terete ; dorsal ones longer and stouter ; in the basal joints of the arms, the uppermost spine is about one and a half times, and the lowest one about two-thirds, as long as the corresponding arm joint. Colour in alcohol yellowish brown. In younger specimens, the skin of the disk and arms is very thick, and the buccal pores are often indistinct. This species is apparently near to 0. brachygnaiha, especially in the disk coverings. But a careful comparison will show that, the present species is more closely allied to 0. asaphcs than to 0. brachygnaiha, as shown in the foregoing key. Ophiosemnotes, g- nov. Disk high and convex, covered by a cereous skin, which obscures the underlying stout scales ; beset with granules or stumpy tubercles. Radial shields exposed, rounded or triangular, more or less separated from each other. Three or four oral papillœ on either side of each jaw. Teeth in a single vertical series. Dental 136 ART. 2. — H. MATSUMOTO : papillae absent. Arms also covered by a cereous skin. Dorsal arm plates entire. Arm spines conical, blunt, stout, opaque. A single conical, blunt tentacle scale to each pore. This new genus includes Ophiolebes tylota Clark, 1911, 0. jMchybactra Claek, 1911, Ophiactis clavigera Ljungman, 1866 (referred to Ophiolebes by Lyman, 1882), Ophiacantha œdidisca Clark, 1911, Ophiolebes cUaphora Clark, 1911, 0. paucispina Clark, 1911, and 0. brevispina Clark, 1911, the first being the genotype. Ophiosemnotes differs from the genuine Ophiolebes chiefly in the short, rounded or triangular, exposed radial shields and in the entire dorsal arm plates. In Ophiolebes as here restricted, the radial shields are long, narrow, bar-like and not exposed, and the dorsal arm plates are divided into two, inner and outer, secondary plates. Ophiosemnotes is very near to Ophiochondrella, but differs from it in the short ventral arm plates, which are separated from one another, and in the single thick, not leaf-like tentacle scale to each pore. In Ophiochondrella the ventral arm plates are quadrangular, as long as the corresponding arm joint and joined to one another, and the tentacle pores are provided with two thin, leaf-like scales. Key to Japanese species of Ophiosemnotes. A — Disk closely covered with fine granules ; radial shields oval, separated from each other ; dorsa] arm plates rather large, in contact with each other in the free basal arm joints, bearing a row of granules along the outer border ; arm spines rather short, upper ones being nearly as long as the corresponding arm joint ; lower arm spines not clavate œdidisca. A A — Disk sparsely bessst Avith clavate tubercles and granules ; radial MONOGIJAPH OF JAPANESE OrHITJEOIDEA. 137 shields triangular, more or less separated or the pairs slightly in contact at the outermost part ; dorsal arm plates small, separated from each other by the lateral arm j)hites, entirely free of granules ; arm spines long, the uppermost one being nearly thrice as long as the corresponding arm joint ; lower arm spines clavate tylota. Ophiosemnotes œdidisca (Clark). OpMacantha œdidisca : 220, fig. 101. One speci- men ; Albatross station 4958, off Hiuga ; 405 fathoms. Off Hinga; 405-578 fath- oms (Claek). Off Omai Zaki ; 505. fathoms (Claek). Off Kii; 544- 545 fathoms {Claek). Though this species ap- parently re- sembles certain species of Ophi- ophthalmus, the high and convex disk, the very Clark, Bull. U.S. Nat. Mus., LXXV, 1911, p. „o'%oo2 Fig. 34. Ophiosemnof es œdidisca. x8. a. From above, h. From below, c. Side view of two arm joints nenv rlisk. 138 AET. 2. — H. MATSUMOTO : small oral region, the slender, skin-covered arms, the narrow ventral arm plates, of which the outer border is slightly notched, the short, stout, skin- covered arm spines and the conical, not leaf-like tentacle scales show that it is only distantly related to that genus. On the other hand, the present species is very near to 0. cliaphora, which appears to me to find its natural place in this germs. Ophiosemnofes tylota (Clark). OpMoJebcs tylota : Clark, Bull. U. 8. Nat. Mus., LXXV, 1911. p. 243, fig. 117. Sea of Japan ; 90-207 fathoms (Clakk). Bering Sea. Family 2. Hemieuryalidae (Verrill, 1899) mihi, 1915. Disk covered with very heavy plates and stout radial shields, the primaries being usually very conspicuous. Interbrachial ventral surfaces very small, with very short genital slits. Oral and adorai shields very thick. Three or four flat oral papillae on either side of each jaw. Teeth triangular, arranged in a single vertical series. Dental papillae absent. Peristomial plates simple, stout, soldered with the oral frames, which are also very stout. Genital plate and scale of the same side of a radius soldered together, very stout. Arms very stout, covered with very heavy arm plates, vertical- ly coiled. Vertebrae stout. Vertebral articulations streptospondyline, articular pegs being very rudimentary or entirely absent. This family includes seven genera, which may be grouped into two subfamilies as follows. Subfamily 1. Ophioi^hondrince (Verrill, 1899) mihi, 1915. — MONOGEAPH OF JAPANESE OPHIUR(^IDEA. 139 Dorsal arm plates entire, without supplementary plates ; five to eight arm spines, moderately long, conical ; no tentacle scales proper, but tlie lowest arm spine may serve as one. Ophiochondrus Lyman, 1869.^^ Ophiomœris Kœhl.^11, 190'i { = OpJdurasesCi.MiK, 1911). Ophiof/yplis Kœhlek, 1905. Subfamily 2. Hemieuryalince Matsumoïo, 1915. — Dorsal arm plates often accompanied by supplementary plates or replaced by a mosaic of secondary i)lates ; three arm spines and one tentacle scale, both being very slK^rt and flat. Sigsbcia Lyman, 1isk entirely free of large, spherical tubercles ; each radial pair of radial shields separated from each other b}- a row of two or three plates ohslricta AA — Large, spherical tubercles present along the outer margins of the radiais and along the adradial borders of the radial shields; radial shields joined in pairs along the whole length prqjecta Ophiomœris obstricta (Ly^fan), Ophioceramis{?) obstricta: Lyman, Bull. Mus. Comp. ZooL, V. 1878, p. 124, PI. VI, figs. 164-166 ; Lyman, Rep. Challenger, V, 1882, p. 26, PL XI, figs. 1-3. OpMomœns obstricta : Kœhler, Kxp. Siboga, XLV, Pt. 1, 1904, p. 17. Ophinrases obstr ictus : Clark, Bull. U. B. Nat. Mus., LXXV, 1911, p. 250, fig, 122. Eastern Sea ; 95-135 fathoms (Clark). Lat. 5^ 42' S., long. 132° 25' E.; 12i) fathoms (Lyman). MONOGRAPH OF JAPANESE OPHIUROIDEA. 141 OpJtiomœris projecta Matsumoto. Ophiomœris projecta : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 65. Two specimens ; off Uki-shima ; 300 fathoms. One specimen ; Uji-shima, Osumi. Diameter of disk 4 mm. Length of arras 13 mm. Width of arms at base 1.5 mm. Disk five-lobed, with concave interbrachial borders, with a very regular rosette of the central, basal and radial plates, surrounded by the radial shields. Cen- tral plate cir- cular, rather large. Eadials rhomboidal, considerably larger than the central plate, separated from each other by the small and oblong basais. The radiais may be in con- tact with the central plate, or separated from it by one p^^ gg ophiomœris projecta. xlO. a. From above. b. From „„ +^^a■r\ XTPW below. c. Side view of three arm joints near disk. 142 ART. 2. — H. MATSUMOTO : small, intervening scales, and bear several very prominent, spherical tubercles along the outer margins. Radial shields very large and stout, triangular, with an acute inner angle, longer than wide, each pair in contact by the outer half of their whole length, and narrowly separated from the next pair by a sunken, oblong plate. The radial shields do not lie in the same plane with the central rosette, but are inclined outwards and interradially, so that the central rosette appears elevated above its surroundings. The radial sliields also bear prominent tubercles along or on their adradial borders, as well as often along their outer margins. Just outside and below the outer edge of the radial shields, there is a very stout and nearly vertical plate, on which the radial shields rest elevated above the general surface of the arm. Interbrachial spaces below very small, covered by a few convex plates. Genital slits small, lying just outside the oral shields. Oral shields large, rhomboidal, wider than long in surface view, but the outer angle is prolonged upwards like a beak. Adorai shields longer than the oral shields, wider without, tapered inwards and meeting each other. Three oral papillae on either side, rather close-set, outermost one quadrangular, largest of all, wider than long ; the next bluntly and the last acutely pointed, both being longer than wide. Four or five teeth, stout, spear- head-shaped, the uppermost one longer and acuter than the rest. Arms stout and sliort, uniformly tapered outwards. Dorsal arm plates triangular, with straight inner sides and two-lobed outer, wider than long, strongly convex, separated from one another by the lateral arm plates, which are very stout and strong- ly flared outwards. First ventral arm plate small, pentagonal, longer than wide, wider within than without, if we except the pointed portion. Second and following triangular, with concave MONOGRAPH OF JAPANESE OPHIUROIDEA. 143 inner sides, onter side convex as a whole, bnt sliglitly concave at the middle, wide and very short. Successive plates separated from one another; distally they become rapidly smaller, and the outer border is then entire. Arm spines conical, acute, short, shorter than the corresponding arm joint ; upper ones longer, lower ones somewhat rough at the end, looking like compound hooks towards the extremity of the arms. They are about seven in number in the first basal joints, but rapidly decrease in number as w^ell as in size, so that, on the eighth lateral plate, they are only four and about half as long as the corresponding joint. The plates of the disk and arms are shagreened. Colour in alchohol : disk gray, arms banded with grayish brown and white. The present species is very closely related to 0. ohslricta (Lyman), but differs from it chiefly in the presence of the prom- inent tubercles on the disk and in the radial shields joining in pairs in the radial line. In the first mentioned character, this species closely resembles OpJiiogj/ptis nodosa Kœhlee, 1905. Order iii. gnathophiurida Matsibioto. Disk covered with fine, imbricating scales, which are very regular in size and arrangement. Radial shields well developed, each having a conspicuous articular pit near the outer end for the reception of a large, ball-like articular condyle of the genital plate. Genital plates usually fixed to the basal vertebrœ. Genital scales short, wide, flat, leaf -like, articulating with the genital plates near the outer end of the latter. Besides, an additional scale is present on the abradial border of the innermost part of each genital slit, being also short, wide, flat, loaf- like and firmly fixed to the oral 144 AKT, 2. H. MATSUMOTO : shield. Peristomial plates entire, or rarely double, nsnally very small. Oral frames very stout, with strongly developed lateral wings as a rule, for the attachment of voluminous masticatory muscles. Oral and dental plates usually very stout, the two plates usually presenting an X- shape in dorsal view. Oral papillae few, often entirely absent. Dental papillaa well developed only in forms without oral papillae. Teeth stout, usually quadrangular, with widened, straight or wavy apical edge. Arms slender, inserted ventrally to the disk, horizontally flexible, or sometimes more or less coiled vertically. Vertebral articulation zygospondyline. Dorsal, lateral and ventral arm plates all well developed. Key to families of GnatJiophiurida. A — Teeth triangular, not very stout ; peristomial plates large ; oral frames entire, without well developed lateral wings ; genital plates free, not fixed to the basal vertebrae ; oral papillse present ; no dental papillae AmpliilepididcB. AA — Teeth quadrangular, very stout ; peristomial plates small ; oral frames with well developed lateral wings ; genital plates firmly fixed to the basal vertebrae. a — Oral papiUse present ; no vertical clump of dental papiUae , AmpJiiuridœ. aa — Oral papillae absent ; dental papillae weU developed, forming a vertical clump at the apex of each jaw Ophiotrichîdœ. Family 1. Amphilepididae Matsumoto, 1915. Disk covered with fine, imbricating scales. Radial sliields Bometimes very rudimentary but often well developed, with a MOXOGRArH OF JAPANESE OPHIUROIDEA. 145 conspicuous articular pit near the outer end for the reception of a large, ball-like articular condyle of the genital plate, which is entirely free from the ]3asal vertebr99. Genital scales short, wide, flat, leaf-like. Peristomial plates entire, very large. Oral frames not very stout, without well developed lateral wings. Oral plates long and slender in interior view. Dental plates very small, or absent. Teeth and oral papillsG present ; but no dental papilla?. Arms slender, inserted ventrally to the disk, being only horizontally flexible. Dorsal, lateral and ventral arm plates all well developed. Three to five arm spines, conical or cylindrical, with smooth surface. One or two tentacle scales. Vertebra} slender, with zygospondyline articulation ; in the distal part of the arms, they are often incompletely divided into halves by a longitudinal series of pores. This family includes three genera, which may be grouped as follows. I. Second oral tentacle pores opening more or less outside the oral slits, so that the latter are gaping. Amphilepis Ljungman, 1866. II. Second oral tentacle pores opening entirely within the oral slits, so that the latter are closed by the oral papillae. Amphiactis Matsumoto, 1915. Ophiochytra Lyäian, 1880. Though the present family much resembles the next family, Amphîurldœ, in external features, the internal structures are rather Lœmophiuridan, with large peristomial plates and entire oral frames. The articulation of the radial shield and genital plate, as well as the flat, thin genital scales, however, gives ground for placing tliis family in the present order. In short, the present family is the 146 ART. 2. — H. MATSUMOTO: least specialised among the Gnathophiurida, and stands next to the Laemophinrida. In Amphiactis umbonata Matsumoto, the genital plate and radial shield articulate with each other by means of a hemispherical articular condyle of the former and a large articular socket of the latter, the former being entirely free from the basal vertebra?. The genital scales are flat and thin, articulating with the genital plates near the outer end. The peristomial plates are simple and very large. The oral frames are small and entire, without well developed lateral wings. The oral plates are long and slender in internal view. The dental plates are absent. According to Lyman, the internal structures of Amphilepis norvegica Ljungman, 1860, appear to be essentially similar to those of the above mentioned species, save that the peristomial plates are somewhat smaller and the dental plates are present. A'inphilepis tenuis Lyman. Ampldlepis tenuis: Lyman, Bull. Mus. Comp. Zool., VI, Pt. 2, 1879, p. 35, PI. XYI, figs. 432-434; Lyman, Rep. Clialleuger, V, 1882, p. 151. Southwards from Boshii ; 1,875 fathoms (Lyiman). Amphiactis Matsumoto, 1915. Disk covered with imbricating scales, radial shields moderate- ly large. Four or five oral papilla? on either side, unequal in size, arranged almost in a continuous series. Teeth arranged in a single vertical row. Peristomial plates entire, very large ; oral frames slender and without lateral wings. Arms long and slender, only horizontally flexible. Distal vertebrae often imperfectly divided into IMONOGRATH OF JAPANESE OPHIUROIDEA. 147 halves by a series of pores. Three to five arm spmes. One or two tentacle scales to each pore. This genus includes Amphmra duplicata Lyman, 1875, A, canescens Lyman, 1879, A. patula Lyman, 1879, A. partita Kœhlee, 1897, Ophiadis cUssidens Kœhlek, 1904, and 0. parata Kœhlee, 1904, besides the genotype, Amphiactis umbonata Matsumoto, 1915. The representatives of the present genus were formerly referred to Amphiura by Lyî^ian, and then to Ophiadis by Lütken & MoETENSEN. AmpJiiadis differs from Amphioplus in the absence of the paired infradental papillae, and from Ophiadis in having numerous oral papilloe arranged in a continuous series, so that the oral slits are entirely closed. Further, the internal structures are entirely different from those of the Amphiuridœ. Again, the present genus diflers from Ophiochytra in the well developed radial shields. Arnphiactis umbonata Matsumoto. Amphiactis umbonata : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 67. Two specimens ; off Misaki, Sagami Sea. Diameter of disk 7 mm. Length of arms 30 mm. Width of arms at base 1.2 mm. Disk circular, flat, covered with rather coarse and irregular scales. Central plate large, circular, surrounded by ten small scales, corresponding in position to the infrabasals and basais. The five radiais large, larger than the central plate, with strongly curved abcentral border, which almost forms a semicircle. The central and radiais have each a small but distinct central or subcentral 148 ART. 2. — H. IMATSUMOTO boss. The second radiais and the first to third interradials may also be distinguished, and are large and prominent. Thus, the disk squamation is rather similar to that of Opliiozona. Eadial shields comparatively small, oblong ovate, about two-fifths as long as the disk radius, twice as long as wide, wider without, more convex abradially than adradially, separated from each other by a row of three or four plates, of which the inner ones are larger. In each interradial space, there are five to seven ii'regularly radiating rows of scales. Interbracliial ventral surfaces covered with coarse, irregular, imbricating scales. Genital slits long, near- ly reaching the margin of the disk. Genital scales invisible. Oral shields small, rhomboidal, with acute innei' and rounded lateral and outer angles, and shghtly concave inner sides. Adorai shields wider without, nearly or quite meeting within. Four oral papilla) on either side of the oral angle, inner ones smaller and more acute. Deep in the oral slits, on either side of the oral angle, there occurs one long, acute, needle-like papilla. Five teeth, all blunt, except the uppermost one, wliicli is pointed. Five arms, slender, flattened, uniformly tapered toward the extremity. Dorsal arm plates large, fan-shaped, twice as wide as long ; inner sides slightly convex and forming an obtuse angle Fig. 36. Amphiartls wnhonata. x8. a. From above. J>. From below, c. Side view of two firm joints near disk. MONOGRAPH OF JAPANESE OPHIUÜOIDEA. 149 within, outer side decidedly convex, outer angles rounded ; succes- sive plates separated by the lateral arm plates, except the basal two or tlii-ee. Lateral arm plates low, not very prominent. First ventral arm plate small, divided into two pieces, of which the inner one is triangular and the other quadrangular. Those beyond large and liexagonal, except the second plate which is pentagonal ; much wider than long, widest at the outer lateral angles, with concave lateral sides and slightly convex inner and outer sides, swollen along the outer borders and especially at the outer angle, so that the arms appear keeled along the ventral median line. Three arm spines, subequal, about as long as the corresponding arm joint, but the uppermost one is slightly longer ; cylindrical, tapered and blunt. Two flat, oval tentacle scales to each pore, but sometimes tliree to the first. Colour in alcoliol white. The present species may be easily distinguished from all the other species of Amphiactis by the Ojjhiozoiia -like squamation, and especially by the presence of the umbonated primary plates. Family 2. Amphiuridae (Ljungman, 1867) mihi, 1915. Disk covered with fine, imbricating scales, or rarely by a naked skin, sometimes beset with minute spines. Eadial shields well developed, with a conspicuous articular socket on the ventral surface near the outer end, fitting to the large, ball-like articular condyle of the genital plate. Genital plates firmly fixed to the basal vertebrae. Genital scales short, wide, flat, articulating with the genital plates near the outer end of the latter. Besides, there occur a pair of short, flat scales, just outside each oral shield, supporting the proximal abradial border of the genital shts. Peristomial plates very small, usually entire. Oral frames very 150 ART. 2. — H. MATSUMOTO : stout, with well developed lateral wings for the attachment of voluminous masticatory muscles. The oral and dental plates are very stout in internal view, the two presenting an X- shape. One to six oral papilte on either side, the innermost one being often infradental. Teeth very stout, widened, squarish, with wavy or notched cutting ends. Arras inserted ventrally to the disk, horizontally flexible, or rarely capable of coiling vertically. Dorsal side of the vertebrae entire, not strongly notched inwards. Vertebral articulation zygospondyline, the articular peg being always present. Arm spines moderately long, conical, stout, opaque. Usually one or two leaf-like tentacle scales, sometimes none. The present family includes eighteen genera, which may be grouped into two subfamilies as follows. Subfamily 1. Ophiactinince Matsumoto, 1915 : — No paired infradental papillae. Ophiactis Lütken, 1856. Hemipholis (Agassiz) Lyman, 1865. Ophiopus Ljungman, 1866. Ophiopholis Müller & Teoschel, 1842. Subfamily 2. Amphiurince mihi, 1915 : — Paired infradental papillae present. I. Four or more oral papillae on either side, the outermost one or two arising from the adorai shield ; one additional papilla is present just outside and above the infradental one ; oral shts closed by the oral papillae. Ampliioplus Veerill, 1899. Amphichilus, nov. Amphiacantha, nov. Ampliilimna Yeerill, 1899. II. Three, rarely four, oral papillae on either side, none MONOGRAPH or JAPANESE OPHIUKOIDEA. 151 arising entirely from the adorai shield ; no additional papilla just outside and above the infradental one, which therefore is the highest in position of all ; oral slits closed by the oral papillae. Ampliioclia Verrill, 1899. Ophiophragmus Lilian, 1865. Ophiocnida Lyman, 1865. III. Three oral papillae on either side, the outermost one being very large and operculiform ; no additional papilla just outside and above the infradental one, which therefore is the highest in position of all ; oral slits closed by the oral papillae. Amphipliolls Ljuxgman, 1867. Ophiostigma Lütken, 1856. IV. Two, sometimes three, oral papilla3 on either side, the outermost one or two arising from the adorai shield ; one additional papilla is present just outside and above the infradental one ; oral slits gaping. Amphiura Forbes, 1842 (= Ophionephthys Lütken, 1869). Ophionema Lütken, 1869. Ctenamphiura Verrill, 1899. Par amphiura Kœhler, 1895. Ophlocentras Ljungman, 1866 {~ Amphiocnida Verrill, 1899). I [have dissected several specimens of Ophiactis pteropoma Clark, Hemipholis elongata (Say), OphiophoUs aculeata (Linnf), AmpMoplus ancistrotm (Clark), AmpJiiacantha acanthina (Clark), A. dividua, no v., Ophiophragmus japonicus Matsumoto, AmphiphoUs kochii Lütken, Amphiura korece Duncan, A. trachydlsca Clark, A. 152 AHT. 2. — H. jNIATSUMOTO : vacUcoIa Matsumoto, Ophiocentrus vert'icillalus (Döderlein), &c., and am strongly impressed by the uniformity of the internal structures of this family. The oral fi'ames are especially stout, with very well developed lateral wings for the attachment of voluminous masticatory muscles ; those of Ophiactis pteropoma are less stout than those of tlie repi'esentatives of the other genera. The oral and dental plates are very stout and more or less X- shaped in internal view. The peristomial phites are rather or very small and usually entire, save in the representatives of Amphiura and Ophiocentrus, in w^hich they are double or rarely triple ; they are comparatively large in Opliiopholis aculeata, Amphioplus auclstrotus, Amphiura Iwrecc, A. trachydisca and Ophiocentrus vertlcillatus, but exceedingly small in Hemlpholis elongata, Ophlophragmus japonlcus and Ampliiura vadicola. The genital plates are firmly fixed to the basal vertebrae, save in Ampldacantlia dicidua, nov., in whicli they are free from the basal vertebrae. This peculiarity of the species just mentioned is probably due to the fact that, the single type specimen is very small and young. The genital plates of Ophio- centrus verticillatus are somewhat unusual in position, lying entirely dorsal, but not lateral, to the basal vertebra\ As to the genital scales and the articulation of the genital plate and radial shield, the statement in tlie diagnosis of the present order holds true for all the representatives of the genera thus far studied by myself. The dorsal side of the vertebrae is usually entire, without any conspicuous notch inwards ; and only in Ophiocentrus verticillatus is it distinctly notched inwards and more or less Y-shaped, rather reminding us of that of the next family. According to MoRïENSEN, Ophlopus arcticus Ljungman has very rudimentary genital bursae, which are represented merely by the creases between the arm bases and the interbrachial ventral surfaces, and the genera- MONOGEAPH or JvlPANESE OPHIÜEOIDEA. 153 tive glands are covered by a sac-like membrane, which contains very fine and delicate scales when viewed under the microscope. Key to Japanese genera of Ox)lnactinmœ, A — Oral angles not exceedingly short ; oral papillaB arising from the oral plate ; dorsal arm plates entire, without supplementary ones . . Oiihiadis, AA — Oral angles exceedingly short ; oral papillœ arising from the adorai sliield ; dorsal arm plates hemmed by a row of supplementary ones along the outer, and often also lateral, sides OpJiiopJioIis. Key to Japanese species of Ophiactis, A — Three arm spines ; disk nearly or entirely free of spines. a— Fire arms; oral shields much wider than long hrachygenys. aa — Six arms ; reproducing by schizogony ; oral shields about as wide as long. b- — Eadial shields separated from each other by a row of three scales ; uppermost arm spine the longest pteropoma. hh — Kadial sliields separated from each other by a single wedge-shaped scale, except at the outer ends, where they are in contact; middle arm spine the longest dyscrita. AA — Four or more arm spines ; disk beset with a number of spines. c — Four arm spines. d — Five arms ; disk scales fine ; radial shields sej^arated from each other by a row of four scales offinis. dd — Six arms ; disk scales coarse ; radial shields partially joined in pairs macrolepidota, cc — Five or more arm spines ; six arms ; reproducing by schizogony. € — Five arm spines in free basal arm joints ; dorsal arm plates exceedingly wide, about thrice as wide as long; single oral papilla on either side modesta. 154 ART. 2. — H. MATSUMOTO : ee — Six arm spines in free basal arm joints ; dorsal arm plates less than, or nearly, twice as wide as long; two or three oral papillae on either side. y — ^Adorai shields meeting with each other both in the radial and interradial lines ; interbrachial ventral sm-faces covered with fine scales savigiujL ff — Adorai shields not meeting with each other either in the radial or interradial lines ; interbrachial ventral surfaces covered by a naked skin gymnocliora. Opliiactis hracJiygenys Clark. OpUactis hrachygenys : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 135, fig. 51. Off Manazaru Zaki, Sagami Sea; 120-265 fathoms (Clark). Off Hiuga; 437-720 fathoms (Claek). Ophiactis pteropoma Clark. OpUactis pteropoma ': Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 134, fig. 50. One specimen; Misaki. Sea of Japan ; 195 fathoms (Clark). Off Kii ; 191-253 fathoms (Clark). This species appears to me to be very near to 0. profondi LüTKEN à MoRTENSEN, 1899, differing from it merely in the arm spines being unequal instead of being subequal. Ophiactis dyscrita Clark. Opliiactis dyscrita : Clark, Bull. Ü. S. Nat. Mus,, LXXV, 1911, p. 137, fig. 52. A , Six specimens, clinging to a hexactinellid sponge ; off Oshima, Sagami Sea ; 75-85 fathoms. MONOGEArH OF JAPANESE OPHIÜROIDEA. 155 Kagosliima Gulf; 58 fathoms (Clark). The largest one of ray specimens measures only 2 mm. across the disk and 10 mm. in the arm length. In some arm joints, the uppermost one of the three arm spines is longer than the middle one, which is usually the longest. The colour in alcohol of these specimens is not gray, but light green. Ophiactis ajfi,nis Duncan. OpMadis affinis : Duncan, Journ. Linn. Soc. London, Xr\^ 1879, p. 469, PI. X, fig. 23, PL XI, fig. 24; Lyman, Ptep. ChaUenger, V, 1882, p. 115 ; Kœhlee, BuU. Sei. Fr. Belg., XXXI, 1898, p. 71 ; Kœhleiî, Res. Exp. Siboga, XLY, Pt. 2, 1905, p. 26. Korean seas (Duncan). Indian Ocean and Malaysian waters (Kœhlee). Ophiactis macrolepidota Marktanner. Ophiactis macrolepidota : Marktanner-Turneretscher, Ann. K. K. Naturliist. Hofmus. Wien, H, 1887, p. 298, PI. Xn, figs. 12 & 13 ; Döder- LEiN, Denkschi'. Med. Nat. Ges. Jena, YIEI, 1898, p. 484, PI. XXXVH, figs. 1 & la. One specimen ; Uraga Channel. Three specimens ; Asami Bay, Tsushima. Sidney (Maektanner-Tueneretschee). Amboina (Dödeelein). The specimens at hand were all found adhering to Ophiothrix koreana. The largest one is 1.8 mm. in the disk diameter and 7 mm. in the arm length. Unlike Maektanner-Turneretscher's type, imbricating scales are visible in the interbrachial ventral surfaces in all the specimens. The disk is yellowish and the arms are light greenish above and light yellowish below in alcohol. 156 ART. 2. — H. MATSÜMOTO Pig. 37. Ophiadis macrokpkloia. x40. «. From above. 7*. From below, c. Side "view of two arm joints near dislc. OpJiiactis tnodesta Brock. Ophiactis modesta : Brock, Zeitscbr. wiss. Zool., XLYII, 1888, p. 482 ; DÖDEßLEiN, Semon - Zool. Forscliuugsr. in Austr. u. Malay. Ai'chip., 1896, p. 285, PL XIV, fig. 1, PI. XV, fig. 5; Kœhler, Mem. Soc. Zool. Fr,, XVU, 1904, p. 63, figs. 10 & 11. Numerous specimens ; Misaki. Amboina. The largest one of the specimens at hand is 4 mm. in the MONOGRAPH OF JAPANESE OPHIUROIDEA. 157 disk diameter and 20 mm. in the arm length. With the excep- tion of three which are five- armed, they are all six-armed. One of the five-armed specimens has three larger and two smaller arms, while another has tw^o larger and three smaller ones. The ventral arm plates have each an hour-glass shaped depression on the surface, which seems to me to be a dis- tinctive character of this species, though apparently unnoticed by previous authors. In Dödeklein's photograph, this de- pression is faintly visible. Colour in al- cohol : dull green ; disk with two or three patches of very dark green ; arms banded, two darker joints alternating with four or fi\'e lighter ones ; each dorsal arm plate with three yellowish spots on the outer margin. Fig. 38. OiMadis moclesta. x20. a. From above. ?;. From below, c. Side view of three arm joints near disk. 158 ART. 2. — H. MATSUMOTO : Ojyhiactis savignyi (Müller k Troschel). Ophiolepis savignyi: Müller & Troschel, 8ys. Ast., 1842, p. 95. OpMölepis sex-radia : Grübe, Wieg. Arch. Naturg., 1857, p. 343.'^ Ophiactis sen-radia : Lütken, Addit. ad Hist., IE, 1853, p. 126 ; Lyman, Hl. Cat. Mus. Comp. ZooL, I, 1865, p. 115; Clark, Ann. New York Acad., XI, 1893, p. 412^>; Kœhler, Bull. Sei. Fr. Belg., XXXI, 1898, p. 72. Ophiactis Tcrebsii : Lütken, loc. cit.; Lyman, loc. cit., p. Ill ; Ljungman, Ofv. K. Vet. Akad. Förh., 1866, p. 323'^; Verrill, Transact. Connect. Acad. , I, 1867, p. 265 ; Duncan, Journ. Linn. Soc. London, XIV, 1879, p. 465 ; Ljungman, loc. cit., 1871, p. 627 ; Rathbün, Transact. Connect. Acad., V, 1880, p. 153 ; Verrill, Bull. Lab. Nat. Hist. Iowa, V, 1899, p. 34^ ); Clark, Bun. U. S. Fish Comm., H, 1901, p. 246. OpJiiactis reirüiardti : Lütken, loc. cit., p. 161, PI. Ill, fig. 7. OpUadis savignyi : Lyman, Rep. Challenger, V, 1882, p. 115 ; Ludwig, Mem. Couronn. Sav. Acad. Belg., XLTV^, 1882, p. 14; Lütken & Morten- sen, Mem. Mus. Comp. ZooL, XXIH, 1899, p. 140 ; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 26 ; Kœhler, Bull. U. S. Nat. Mus., LXXXIV, 1914, p. 41. Two specimens ; Koajiro, Misaki. Numerous specimens (be- longing to the First High School); Misaki. Korean Seas (Duncan). Malaysian waters. Australia. Sandwich Is. Gulf of Cali- fornia. West Indies. The larger one of the two specimens mentioned first is 3.5 mm. in the disk diameter and 10 mm. in the arm length. The smaller one is 2 mm. across the disk. Both are six-armed, and have no distinct central plate. In the larger specimen, the dorsal arm plates are often divided into two or three pieces, and are scarcely lobed on the outer side; the colour is yellowish brown in alcohol, and faintly banded on the arm near the extremity. In the 1) These papers were not seen by me. MONOGRAPH OF JAPANESE OPHIUROIDEA. 159 smaller one, the colour is vivid green in alcohol ; the arms have darker and lighter bands of the same colour ; and each dorsal arm plate has three wliite spots on the outer margin, which are especially conspicuous in the dark bands. It seems that, in this species, the colour changes in alcohol from green to brown in the course of time. Ophiactis gymnocJiora Clark. OpJdadis gymnocliora : Clark, BuU. Ü. S. Nat. Mus., LXXV, 1911, p. 139, %. 54. Tanega-shima (Clakk). Fig. 39. Ophiactis saviçpiyi. x 16. a. From above, b. From below, c. Side view of two arm joints near disk. Key to Japanese species and varieties of Opliioi^UoUs, A — One very large supplementary plate present on either side of each dorsal arm plate, besides the smaller supplementary ones, .mirahilis. AÄ — No especially large supplementary dorsal arm plates, though there is present a row of small supplementary plates along the outer, and often also lateral, borders of each dorsal arm plate. a — Supplementary dorsal arm plates occurring along both the outer and lateral borders of the primary ones ; oral papillae short and stout . . aculeafa. &— Arm spines short, stout, flattened, blunt; radial shields mostly or entirely covered with granules or spines typical aculeata. 160 AßT. 2. — H. MATSUMOTO : hh — Arm spines long, slender, conical, ratlier pointed ; radial shields mostly or entirely naked var. japonica. aa — Supplementary dorsal arm jplf^'tes occurring only along the outer border of the primary ones ; oral papiUse long and slender .... hracliyactis. OphiopJiolis mirabilis (Duncan). Ophioleiyis mirabilis : Duncan, Jouru. Linn. 8oc. London, XIV, 1879, p. 460, PI. IX, fig. 12, PI. X, figs. 13 c^ 14. Ophiopholis mirabilis : Lyman, Bull. Mtis. Comp. Zool., YI, 1879, ]). 43 ; Lyman, Rep. Challenger, Y, 1882, p. 115 ; Kœhlee, Res. Camp. Sei. Monaco, XXXIY, 1909, p. 168; Clark, Bull. U.S. Nat. Mus., LXXY, 1911, p. 117, fig. 43. Numerous specimens ; o& Misaki. Numerous specimens (be- longing to Mr. B. Saito, Imperial Museum at Uyeno,Tokyo); Aomori Bay. Four specimens ; off Hokkaido, Okhotsk Sea. Korea (Duncan). L'^raga Channel ; 58 fathoms (Claek). OffDoumiki Saki ; 01 fathoms (Claek). Off Ando Zaki ; 76-79 fathoms (Claek). Off Kinkwasan ; 78 fathoms (Claek). Hakodate Bay; 11.5-22 fathoms (Claek). The largest of these specimens is 10 mm. in the disk diameter and 40 mm. in the arm length. Most specimens are brownish gray in alcohol, variegated or spotted on the disk and banded on the arms with dark gray, and often patched on the disk with white. Some other specimens are yellowish brown, variegated and Imnded with dark brown, while still others are light brown, not variegated or banded. One specimen is entirely white. Tlie relative size of the dorsal arm plates and the two large supple- mentary plates is also variable, ranging from 2:1 to 1:1. MONOGRAPH OF JAPANESE OPHIUKOIDEA. 161 OiyJiiopholis aculeata (Linné). Astcrias acîdeaia: Linné, Sys. Nai, 1767, p. 1101. OpJdopholis aculeata : Gray, Kad. Auim. Brit. Mus., 1848, p. 25^^; Lütkex, Addit, ad Hist. Opli., I, 1861, p. 60, PL II, figs. 15 & 16 ; Lyman, Rep. Chal- lenger, V, 1882, p. 112, PI. XLVI, fig. 6 ; Ludwig, Zool. Jahrb., I, 1885, p. 285; Bell, Cat. Brit. Mus. Echinod., 1892, p. 125; Grieg, Eergens Mus. Aarbog, 1892 (1893), No. 3, p. 9; Greeg, Fauna Ai'ctica, I, 1900, p. 264 ; Michailoysky, Ann. Mus. Pétersbourg, 1902, p. 494; Grieg, Ber- gens Mus. Aarbog, 1902, p. 13 ; Nichols, Proc. R. Irish Acad., XXIX, 1903, p. 257 ; Norman, Ann. Mag. Nat. Hist., Ser. 7, XH, 1903, p. 407 ; Clark, Bull. U. S. Fish Comm., 1904, p. 559, PL V, figs. 24-27, PL VH, figs. 41 & 52 ; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, p. 298 ; Kœhler, Kés. Camp. Sei. Monaco, XXXIV, 1909, p. 167 ; Clark, Bull. U. S. Nat. Mus., 1911, p. 128, fig. 48 ; Nichols, Proc. R. Irish Acad., XXXI, 1911, p. 57; Kœhler, Bull. U.S. Nat. Mus., LXXXrV^ 1914, p. 38. Eleven specimens (belonging to Mr. H. Asano, Imperial Bureau of Fishery); off Kitami ; 60 fathoms. 4^ Fig. 40. Ophiopholis mirabilis, x 6^. a. From above. h. From below, c. Arm spines of one side of an arm joint near disk. d. Arm spines of one side of an arm joint near the extremity. 1) This paper was not seen by me. 162 ART. 2. — H. MATSUMOTO Sea of Japan ; 59-190 fathoms (Clark). Gulf of Tartary ; 318 fathoms (Clark). Okhotsk Sea; 64-73 fathom.s (Clark). North Pacific. Arctic Ocean. North Atlantic. Opiiiopholis aculeata var. japonicci (Lyman). OpMopJioIis japonica : Lyman, Bull. Mus. Comp., Zool., YI, Pt. 2, 1879, p. 42, PI. Xin, figs. 374-376 ; Lyman, Piep. Challenger, V, 1882, p. Ill, PI. XXm, figs. 13-1Ö ; Kœhler, Pve's. Camp. Sei. Monaco, XXXIV, 1909, p. 1G8. OpMopholis aculeata var. japonica : Clark, Bull. Ü. S. Nat. Mus,, LXXV, 1911, p. 123, fig. 47. Numerous speci- mens ; off Misaki, Saga- mi Sea. Three speci- mens ; off Hokkaido, Okhotsk Sea. One specimen ; Albatross station 4784. One specimen ; Albatross station 4791. Off Omai Zaki, Yenshii Sea; 475- 565 fathoms (Clark, Lyman). Sagami Sea ; 369-775 fathoms (Clark, Lïiian), Uraga channel ; 58-302 fathoms-(CLARK). Off Kinkwasan; 82 fathoms (Clark). Off Port Arari; 400-726 Fig. •41. OpMopholis aculeata var. japonica. x6. a. From above. 5. From below, c. Arm spines of one side of an arm joint near disk. d. Arm spines of one side of an arm joint near the extremity MONOGROrH OF JAPANESE OPHIUROIDEA. 1G3 fathoms (Clark). Off Kii ; 244-649 fathoms (Clark). Off southern Hokkaido; 175-349 fathoms (Clark). Off Korea, Sea of Japan; 163 fathoms (Clark). Sea of Japan; 59-428 fathoms (Clark). G-ulf of Tartaiy; 318 fathoms (Clark). Off Saghahn ; 21-32 fathoms (Clark). Okhotsk Sea; 64-100 fathoms (Clark), Yezo Strait; 86 fathoms (Clark). Off Simushir Is.; 229 fathoms (Ciark). Bering Sea. Alaska. Kamchatka. Ophiopholis hrachy actis Claek. O'pUopliolis hrachyactis : Clark, Bull. U. 8. Nat. Mus., LXXV, 1911, p. 117, fig. 44. Numerous specimens ; off Misaki. Numerous specimens ; off Uki-shima, Uraga Channel. One specimen, attached to Sijncdlactes ishikawai Mitsukuri ; Albatross station 5092. Off Ando Zaki; 79 fathoms (Clark). Off Kii; 191 fathoms (Clark). Uraga Channel ; 88-197 fathoms (Clark). Off Manazuru Zaki, Sagami Sea; 153 fathoms (Clark). Suruga Gulf; 108-131 fathoms (Clark). Eastern Sea; 103-361 fathoms (Clark). The largest of these specimens is 11 mm. in the disk diameter and 66 mm. in the arm length ; a medium-sized one is 9 mm. in the disk diameter and 34 mm. in the arm length ; and the smallest 4 mm. in the disk diameter and 15mm. in the arm length. Thus, the arms are in none of my specimens so short as Clark states, and indeed in the first specimen, they are about six times as long as the disk diameter. It appears to me that this species almost passes on into 0. aculeata var. japonica. Some specimens of the latter, especially of Misaki and vicinity, frequently lack the disk spines and are less granulated on the disk, so that they look 164 AKT, 2.— H. MATSUMOTO almost exactly like 0. brachyactis, so far as the disk is concerned ; others again have very short and stout arm spines like those of the typical hrachyadis, while the supplementary plates on the sides of the dorsal arm plates are like those of the typical japonica. In the neighbourhood of jNIisaki, the northern japonica and the southern brachyactis occur side by side, and it is very possible that natural hybridisation may occur. Both these species are found clinging to gorge - naceans and hexactinellids, and are very often found together on the same piece of those animals. Fig. 42. OxMoplioUs hrachyadis. x 8. a. From above. h. From below, c. Arm si^ines of one side of an arm joint near disk. d. Arm spines of one side of an arm joint near the extremity. Revision of the genera of A^nphiurlnœ, The forms which I now bring together under the Amphiurinœ, have hitherto been referred to several genera chiefly on the ground of the character of the disk coverings, such as the presence or absence of scales, complete or localised squamation, the presence or absence of disk spines, &c. But I believe that, such a classi- MONOGRAPH OF JAPAÎvîESE OPHIUROIDEA. 165 fication is very superficial, being based upon characters of merely secondary importance, as may be seen at once on looking over the whole series of this subfamily. The creation of Hemilepis and Opliiopellis by Ljungman and of Opliionephthys by Lütken ; the referring of those species which I now place in AmjiJiiadis, to Amphiura by Lyman ; the referring of those genera which I now place in the Ophiochitonidœ, to the Amplnuridœ by many authors ; the referring of a genuine Amphiura to HemiplioJis by Duncan, etc. are, in my opinion, some of the more notable errors which have arisen from this superficial classification. Vekeill has divided Amphiura and Ophiocnida in a wide sense respectively into five and three distinct genera by the character of the oral papilla?. I am obliged to adopt his subdivisions as a principle, because I believe that, it is scarcely possible otherwise to elucidate the interrelationships of the genera of the Amphiurince. We see that, in almost all the genera of the present subfamily, the oral papilla? are quite similar in form and arrangement to those of any one of Veeeill's subdivisions of Amphiura in a wide sense. Thus, Ophionephthys, Ophionema, Paramphiura, Ophiocentrus, and Veeeill's Amphiocnida correspond in the character mentioned to Amphiura in A^eeeill's sense ; Ophiophragums, as well as Ophio- cnida in Veeeill's sense, to Veeeill's Amphiodia ; Ophiostigma to Amphipholis in Veeeill's sense ; and Veeeili/s Amphilimna to Veeeill's Amphioplus. I look upon these relations to be of primary importance, being evidently more fundamental than the characters of the disk coverings. Granting this, I believe there are two ways open for us ; one would be to unite most genera of the present subfamily into a single genus, and the other to subdivide Amphiura and Ophiocnida in a wide sense into many genera. The first way appears to me 166 ART. 2. — H. MATSUMOTO : to be less adapted to the purpose of bringing out the interrelation- ships of the different subdivisions of this subfamily, which is very- extensive and very rich in species and species groups. Therefore, I am obliged to adopt the second way. We may also ask if Veeeill's subdivisions of Amphiura s. ext. are sharply distinguishable from one another or not. My answer is affirmative. Unfortunately, Amphiura s. str. and Amphiodia, or Amphiodia and Amphioplus are frequently confused by certain authors. My method of distinguishing them from each other is as follows. I distinguish three groups of oral papillae in the present subfamily : the first group is infi-adental, arising from the dental plato^-*; the second arises from the oral plate ; the third arises from the adorai shield. When a papilla arises partly from the oral plate and partly from the adorai shield, it is referred to the second group. Now, let +1 denote the presence of the first group, —I its absence, +11 the presence of the second group, &c. Then, the oral papilla3 of Veeeill's subdivisions of Amphiura s. ext. may be shown by formulae as follows. Amphioplus- gvow^ : + 1 + II + III. Amjjhiodia-gro'op : + 1 + II — III. Amphipholis- gvoM\) : +1 +11 — III."^ Amjyhiur a- group : +1 — 11 + III, or rarely +1 +11 +III. In both the Amphioplas- and Amiyhiura-group, there is present an additional papilla just outside and above the infi*adental one, so that the latter is not the highest in position of all the papillae. 1) The infradental papillic arise from the oral plates, notwithstanding its intimate relation to the dental plate in the adult (H. L. Cjlaek : Growth-changes in Brittle Stars; Publication No. 182 of the Carnegie Inst., Washington, 1914.). 2) The outermost oral papilla of the AmjMpholis-gioup arises from the adorai shield, not- withstanding its partial relation to thel ora plate in the adult (Ci.aek : loc. cit.). MONOGEAPH OF JAPANESE OPHIUEOIDEA. 167 In both the Amphiodia- and Ainphipholis-grou-p, on the contrary, such an additional papilla is entirely absent, and the infradental one is the highest in position. In the Amphioplus-gvoxrg, all the three groups of oral papillae are well developed and form a continuous row, while in the Amphiura-gï:o\\^, the second group of oral papillœ is entirely, — or mostly, as in certain species of Ainphiura, — absent. In the Amphiodia- and AmphiphoHs-growp, the third group of oral papillae is absent, the two subdivisions being distinguished by the different development of the outermost oral papilla. It must be clear from the above exposition, that the Amphiodia- group is not intermediate between the Amphioplus- and Amphiura- group, but the latter are directly interrelated without the intermedia- tion of the first. In my opinion, those species of Amphiiira s. str. with two distal oral papillœ on either side are rather intermediate between Amphiura with a single distal papilla and Amphioplus, instead of being intermediate between the former and Amphiodia. My division of the Amphiurinœ into four groups almost coin- cides with Veekill's subdivision of Amphiura s. ext., but not with Kœhlee's. Most species of Kcehlee's Amphiodia (e.g. those in Res. Siboga Exp.), as well as Amphiodia^ digitula Claek, 1911, appear to me not to be genuine Amphiodia, but to be Amphiura having however two distal oral papillae, of which at least the outermost one arises from the adorai shield. Key to genera of A^mpJiiurince. AmpMophts-gronp :— Oral papiUae +1+11 +111, four or five on either side, forming a continuous row, so that the oral slits are more or 168 AKT. 2, — H. MATSUMOTO : less perfectly closed ; an additional papilla present just outside and above the infradental one. a — Disk entirely free of spines. h — Outermost oral papilla small and not operculiform ; radial shields divergent AmpJiioplus. hh — Outermost oral papilla very large and operculiform ; radial shields more or less perfectly joined Ampldddlus. aa — Disk beset with scattered spines, c — Oral papillie close-set, the outermost one being not very long and spiniform but short and flat ; radial shields divergent ; three to five arm spines ; tentacle scales short, flat, leaf-hke AmpJdacantha. cc — Oral papillae well spaced and conical, the outermost one being very long and spiniform ; radial shields perfectly joined ; six to ten arm spines ; two spiniform tentacle scales, of which the adradial one is very long and slender Ampldlimna. II — AmpJdodla-group : — Oral papillae +1+11—111, three or four on either side, subequal, forming a continuous row, so that the oral slits are more or less perfectly closed ; no additional papilla just outside and above the infradental one. d — Disk free of spines and granules. e — Disk soft, witliout special marginal scales or spines ; radial shields usually divergent Amphiodia. €6 — Disk solid, with a row of special marginal scales or spines, so that the boundary between the dorsal and ventral sides of the disk is very sharp ; radial shields perfectly joined Ophiophragmus. dd — Disk with numerous scattered spines or granules Ophiocmda. Ill — AmpTdpholis-gYoxi]):—Ova\ pa pillas +1 +11 —III, three on either side, the outermost one being very large and operculiform ; oral slits perfectly closed ; no additional papilla just outside and above the infradental one. f — Disk entirely free of spines AmpMpholis. ff — Disk with scattered spines or granules Opidostigma. IV — Ampldnra-gvon-p : — Oral papillae +1—11 +111 or sometimes +1 MoNOGIÎArH OF JAPANESE OrHIUROlDEA. 169 + 11 +III, two or three ou either side, discontinuoiis, the iufradeutal and distal ones being separated from each other by a wide interval ; oral slits gaping ; an additional papilla present just outside and above the infradental one ; radial shields divergent or separated from each other. g — Disk entirely free of spines. h — No supplementary plates between the oral plates and the adorai shields. i — Oral shields not very stout, not in contact with the iîrst lateral arm plates. ./ — Disk entirely or partially covered ^ith fine scales ; radial shields not very narrow and bar-like Amphinra. jj — Disk entirely covered by a smooth naked skin ; radial shields exceedingly narrow and bar-like OpJiionema. it — Oral shields very stout, in contact with the first lateral arm plates ; two distal oral papiUae, of which the outer one is small and spiniform, while the inner one is very large and flat ; ten arm spines Ctenamphiurœ Jih — Two supplementary plates present between the oral plates and the adorai sliields of each interradius ParampJdura. (J g — Disk beset with numerous spines OpJdocentrus. Key to Japanese species of Aniphioplns. A — Two tentacle scales to each pore. a — Three arm spines. ^ — Tentacle scales unusually large, nearty closing the large pores ; ventral arm plates much wider than long ; disk unknown .... megopomus. l>b — Tentacle scales small, not closing the large pores ; ventral arm plates longer than wide ; radial shields slender, partly joined in pairs rhadinohracldus. aa — Four or more arm spines. 170 AET. 2. — H. MATSÜMOTO : c — Four or five arm spines, of which the second lowermost is hook-shaped at the tip and microscopically serrate along the concave side ; radial shields small, divergent ; outer end of each genital plate bearing, just distally to the radial shield, a small projection, which is microscopically thorny at the tip ancistrotus. cc — Six arm spines, one or more middle ones terminating with a minute, glassy cross-bar ; upper arm plates hexagonal, twice as wide as long ; disk unknown hexacanilms. A A — Single tentacle scale to each pore. d — Three arm spines, of which the uppermost one is the longest and the lowest one the shortest; disk scales rather coarse, with very distinct six primaries ; radial shields very large and wide, divergent ; oral shields very small, triangular; ventral arm plates hexagonal, in contact with each other for a short extent ; tentacle scales large and oval cernuus dd — Four or five arm spines. e — Four arm spines, of which the uppermost and the lowest ones are longer than the middle two ; disk scales fine ; radial shields slender, partly joined in pairs ; interbrachial ventral surfaces covered by a naked skin ; oral shields small, rhomboidal ; ventral arm plates quadrangular, widely in contact with each other ; tentacle scales small and crescentic glaucus.. ee — Five arm spines, of which the lowest one is the longest ; disk scales fine ; radial shields long and rather narrow, divergent ; oral shields large, pentagonal, much wider than long ; ventral arm plates quadrangular, much wider than long, widely in contact with each other ; tentacle scales small and oval macraspis. Arnx*liiopliis inegaxmmus Clark. AmpMophiG megapomiis : Clark, BuU. U. S. Nat. Mus., LXXV, 1911, p. 170. Oif Eai, Kitan Channel ; 37 fathoms (Clakk). MONOGRAPH OF JAPANESE OPHIUEOIDEA. 171 Ampliioplus rhadinohvacliius Clark. AmpJdura sp.: Lütken & Mortensen, Mem. Mus. Comp. Zool., XXITE, 1899, p. 158, PI. Xin, figs. 1-3. Ampldoplds rhadinohracliius : Clark, Bull. U. S. Nat. Mus,, LXX"V 1911, p. 169, fig. 74. Niimerons specimens ; Okinosé (a submarine bank), Sagami Sea ; 85 fathoms. Sagami Sea ; 369-405 fathoms (Claek). Suruga Gulf ; 282- 503 fathoms (Clakk). In the larger specimens, the interbrachial ventral surfaces are almost naked, though in the smaller ones they are distinctly squamated, as shown in Claek's figure. Lütken & Moetensen's species, which was not named because the disk was wanting, agrees very well with the present species. Many of my specimens also lack the whole dorsal side of the disk. Amphioplus ancistvotus (Claek). AmpJdodia ancistrota : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 161, fig. 69. Numerous specimens ; Okinosé (a submarine bank), Sagami Sea; 85 fathoms. Two si^ecimens ; Yahagi-gaké, ofFMisaki, Sagami Sea; 310 fathoms. Uraga Channel; 70-197 fathoms (Claek). Off Manazuru Zaki, Sagami Sea; 153 fathoms (Claek). Off Osé Zaki, Suruga Gulf; 45-65 fathoms (Claek). Off Kii, Kitan Channel; 191 fathoms (Claek). Sea of Japan; 61 fathoms (Claek). In my specimens, the radial shields are narrower and less divergent than in Claek's type. The oral shields are rhomboidal and longer than wide, but variable. On either side of the oral 172 ART. 2. — H. MATSUMOTO Fig. 43. Amph'ioplus ancistrotus. a. From above, x 8. h. From below, x 8. c. Ventral view of an oral angle, x 16. d. Arm spines of one side of an arm joint near disk, x 16. angle, there are usually four oral papillœ, of which the outer- most one arises from the adorai shield ; besides, there occurs an additional papilla, placed above the level of the ordinary papilla?, but visible from below just between the infradental and the next papilla. In some speci- mens, there are five ordinary and one additional papillœ, of which the outermost two arise from the adorai shield (fig. c). Thus, the mouth parts of the present species present the characters of Amphioplus, and not of Amphiodia. p. 70. Atnxiliioplus JtexacanfJitis Clakk. AmpMoplus hexacanthus : Clark, Bull. Ü. S. Nrat. Mus., LXXV, 1911, I. Uraga Channel ; 88 fathoms (Clark). Off California ; 50 fathoms (Claek). Amx^Jnoplus cernuus (Lyman). Ampldura cernua : Lyman, Bull. Mus, Comp. Zool., YI, Ft '2, 1879, p. 28, PI. Xn, figs. 323-325; Lyman, Hep. Challenger, Y, 1882. P. 138, PI. XYn, figs. 13-15. MONOGIUPH OF JAPANESE OPHIUKOIDEA. 173 Ampldophis cernua : Verrill, Transact. Connecticut Acad., X, 1899, p. 315. Eastwards from Honshu ; 2,300 fathoms (Lyman). Amphioplus glaiicus (Lyman). Amphmra. glauca : Lyman, Bull. Mus. Comp. Zool,, VI, Pt. 2, 1879, p. 29, PL XII, figs. 326-328; Lyman, Eep. Challenger, V, 1882, p. 139, PI. xvin, figs. 1-3. Ampldoplus glauca : Verkill, Transact. Connecticut Acad., X, 1899, p. 315. Sagami Sea ; 345 fathoms (Lyivian). Yeushu Sea ; 420 fathoms (Lyman). Aitiphiopltis macrasins (Clark). AmpModia macraspis : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 157, fig. 66. Numerous spechnens ; Otaba, Sagami Sea ; 500 fathoms. Off Honshu; 501-749 fathoms (Claek). Suruga Gulf; 60- 270 fathoms (Claek). Sagami Sea; 622 fathoms (Claek). Off Korea; 184 fathoms (Claek). Gulf of Tartary ; 318 fathoms (Claek). Off Washington ; 115 fathoms (Claek). In the larger specimens, the radial shields are rather narrower and the ventral arm plates comparatively wider. As far as I have observed, there is present in all cases a single tentacle scale to each pore. The oral papilla) are very variable. There are present usually four genuine oral papillœ on either side, besides an additional one, which occurs just between the infradental and the distal papilla\ The additional papilla is smaller and 174 ART. 2. — H. MATSÜMOTO : Fig. 44. Ampldoplus macraspis. x 9. a. From below, h. Side view of three arm joints near disk, c. Ventral view of an oral angle, d. Ditto. acuter than the ordinary ones» and is sometimes placed at a liigher level than the distal oral papillaG (fig. d), but very often forms a continuous series with the latter (fig. a). Some- times, there occur only two genuine distal papillaB (fig. c), besides the comparatively large additional one. The outer- most papilla is very small, often so small as to be almost imperceptible ; the next is the largest of all. These two arise from the adorai shield. Sometimes, a single (fig. a) or paired (fig. d.) very small additional infradental papillae are present, just above and between the pair of the ordinary ones. As already mentioned, Amphiodia does not stand intermediate between Ampliiura and Amphioplus, but the last two genera are directly related to each other. In my opinion, the species of Amph'mra with two distal oral papillœ on either side, such as A. cmomala Lyman, A. crassa Kœhlee, A. servata Kœhler, A. grata Kœhler, a. reposlta Kœhleu, A. korece Duncan, A. assimiUs Lüt- ivEN à MoETENSEN, A. cligitula (Claek), t^c. are Amphioplus-Yike) AmpMura, while the present species is a representative of Amphiura- like Amphioplus. The oral angle shown in fig. c more or less reminds us of that of an Amphiura. MONOGKATH OF JAPANESE OPHIUROIDEA. 175 Amphichilus, g. nor. Disk covered with fine, imbricating scales. Radial shields small, narrow, usually joined in pairs. One apical and two distal pairs of oral papillie to each oral angle ; the apical pair are infradental, arising from the dental plate ; the inner pair of the distal ones arise from the oral plate ; the outermost pair are very- large, wide, operculiform, arising ù^om the adorai shield. Besides, an additional papilla occurs just above and between the apical and inner distal papillae, so that the oral shts are perfectly closed. Arms long, slender, flattened, inserted ventrally to the disk, only horizontally flexible. Three arm spines. One or two tentacle scales to each pore. This new genus includes Amphiura dalea Lyman, 1874, and A. intermedia Kœhlee, 1905, besides the genotype, Amphichilus trichoides, nov. Amphichilus differs from Amphioplus in the presence of only three genuine oral papillse, besides an additional one, on either side of each oral angle ; in the very large and operculiform outer- most oral papilla ; and in the usually joined radial shields. It differs fi-om Amphlpholis in the discontinuous arrangement of the genuine oral papillae, and in the presence of the additional oral papilla, which fills up the space between the apical and inner distal papillae. Ainphichilus trichoides , sp. nov. One specimen ; locality unknown, perhaps Sagami Sea. Diameter of disk 6 mm. Length of arms CO mm. Width of arms at base 0.5 mm. Disk circular, rather convex dorsally, covered with very fine. 176 AET. 2. — H. MATSÜMOTO imbricating scales, which are finer outwards and along the sides of the radial shields. Primary plates indistinct. Radial shields long and narrow, acutely pointed within, about one-fifth as long as the disk diameter, meeting with each other for nearly their whole length but parted at the proximal end. Interbrachial ventral surfaces covered also with imbricating scales, which are somewhat finer than those of the dorsal surface. Genital slits long. Oral shields ovate, or rather triangular with per- fectly rounded angles, lateral sides quite and outer side strongly convex ; much longer than wide. Adorai shields triangular, tapered within to a point, where they nearly meet. The apical oral papillae are oval and stout. Two distal oral papillae on either side ; the outer one is the largest of all, much wider than long, flat, thick, has curved free edge, and arises from the adorai shield ; the other longer than wide, flat, thick, with obtuse end. Besides, there occur above the Fig. 4S. Amphlchilus trichoides. x20. a. From above, h From below, c. Side view of three arm joints near disk. MO^OGlîArH OF JArANESE OPHIÜROIDEA. 177 level of the ordinary ones between the apical and tlie next papilla, an acute additional papilla, which together with the others closes the oral slit. Arms long and very slender, flattened. Dorsal arm plates triangular, with much rounded angles, inner sides slightly and outer side strongly convex, wider than long, successive plates shghtly in contact. Lateral arm plates low, joined on the sides, but not meeting above ; below, the plates meet slightly beyond the basal region of the arms. First ventral arm plate very small, pentagonal, longer than wide ; those beyond hexagonal, nearly as long as wide, slightly wider without than within, successive plates a little in contact in the basal joints ; but more distally, they are separated by the lateral plates, and are pentagonal. Three arm spines, subequal, about as long as the corresponding arm joint, conical and obtuse. Two tentacle scales, a larger adradial and a smaller abradial forming nearly a right angle. Colour in alcohol light yellowish gray. This new species differs from A. daleus (Lyman) in the narrower and perfectly joined radial shields, in the smaller adorai shields, in the shape of tlie ventral arm plates and in the presence of two tentacle scales ; and from A. intermedins (Kœhler) in the shape of the radial and oral shields, in the shape of the oral papilLT, especially of the infradental ones, and in the narrower arms. The last species appears to be most nearly allied to the present. Amphiacantha, g. nov. Disk covered with fine, imbricating scales, and scattered spines. Eadial shields usually divergent, sometimes entirely 178 AliT. 2. — H. MATSÜMOTO : separated from each other. Four or five pairs of oral papilla3 to each oral angle, short, close- set ; the apical infradental pair arise from the dental plate, and the outermost pair from the adorai shield. Arms long, slender, flattened, inserted ventrally to the disk, only horizontally flexible. Three to six arm spines. One or two small, leaf-like tentacle scales. This new genus includes Oplnostigma fonnosa Lutken, 1872, Ophiocnida sexradia Duncan, 1889, Amph'iura notacantha Lutken & MoETENSEN, 1899, A. gastracantha Lutken & Moktensen, 1899, Ophiocnida libera Kœhler, 1907, 0, amphacantha McClendon, 1909, Amphilimna pentacantha Clark, 1911, and Amphiacantha dividua, noY., besides the genotype, Amphiopliis acanthinus Claek, 1911. Amphiacantha is practically Amphioplus with disk spines, and differs fi^om Amphilimna Veeeill (non Claek, 1911), in the divergent radial shields, in the close-set oral papilla^ in the not very long, but short and flattened outermost oral papilla, in the fewer arm spines, and in the not long and spiniform, but short and leaf-like tentacle scales. Key to Japanese species of AnipJnacantha, A — Three arm spines ; two tentacle scales ; radial shields not very small, joined or closely set in pairs. a — Disk spines very numerous, very short, l.iluut, present ou both tlie dorsal and ventral sides ; radial shields joined in pairs ; oral shields mucli wider than long; ventral arm plates Lmger than wide formosa. aa — Disk spines not very numerous, very long, slender, acute, present only on the dorsal side ; radial shields divergent, the pairs being in contact only at the outer ends ; oral shields much longer than wide ; ventral arm plates wider than long acanthiua. :\[ON(XiEArH or Japanese ophiukoidea. 179 dA — Four arm spines ; single tentacle scale ; radial sliielcls very small, widel}' separated from each otlier ; disk spines i^ot very mimerous, very short, blunt, present only on the dorsal side d'lvidua. Amphiacantha forniosa (Lütken). Opldostigma fonnosa : Lütken, Oph. Nov., 1872, p. 3, PI. I-II, figs. ^a k, 55 ; Lyman, Rep. Challenger, V, 1882, p. 166 ; Ivehler, Bull. Sei. Fr. Belg., XXXI, 1898, p. 76; Ivœhler, Ees. Exp. Siboga, XLV, Pt. 2, 1905, p. 24, PI. n, fig. 1. Strait of Formosa (Lütken). Indian Ocean and Malay.sian waters. As far as I can judge from Lütken's description and figures, this species is evidently not referable to OphiosUgma, having four oral papilla?, of which the outermost one is very small and rudi- mentary, cjuite as in the next species. Aiupft iacff nth a acantftina (Clark). Ampliioplus acanfMnus: Clark, Bull. U.S. Nat. Mus., LXXY, 1911, p. 168, fig. 75. Two specimens ; Uraga Channcd ; 140 fathoms. Off Tsurugi Saki; 110- 259 fathoms (Claek). Suruga Gulf; 148 fathoms (Claek). The disk scales are fine and re^'ular in one of the Fig. 46. Amphiacantha acanthina. x5. a. From above, h. From below, c. Side view of three arm joints near disk. 180 AIÎT. 2. — H. MATSUMOTO : specimens, but coarse and irregular in the other. The oral papilla) are usually five on either side, but four in some of the oral angles. The outermost papilla is very small. The disk spines are rather more numerous than in Claek's type. Amx>hiacantha dividuaf sp. uov. One specimen ; locality unknown, probably Sagami Sea. Diameter of disk 3 mm. Length of arms 15 mm. Width of arms at base 0.7 mm. Disk subpentagonal, with slightly convex interradial borders, covered with very fine, imbricating scales and scattered spines, which are very short, conical and obtuse. Radial shields very small, subtriangular, with the adradial side longest, about twice as long as wide. Inter- brachial ventral surfaces covered also w^ith imbricating scales, which are somewhat finer than those of the dorsal side ; without any spines. Genital slits long. Oral shields oval, or rather rhomboidal with round- ed angles and convex sides, distinctly longer than wide, wider without than within. Adorai shields triangular, long, rather narrow, tapered in- wards to an acute point, not Fig. 47. Amphhicantha dividua. xl6. a. From fibove. h. From below, c. Side view of two arm joints near disk. MONOGEArH OF JAPANESE OrHIUROIDEA. 181 in contact with each other. Oral plates long and narrow. Six oral papilla} on either side, flat, obtuse, except the outermost one, which has pointed inner end. Teeth quadrangular, stout, with truncate free end. Six arms, of which three are longer and stouter than the others. Dorsal arm plates hexagonal, with rounded outer angles, slightly longer than wide, widest at the lateral angles, in contact with each other. Lateral arm plates low, those of the two sides not meeting above or below. First ventral arm plate very small, pentagonal, longer than wide. Those beyond hexagonal, with concave lateral and notched outer borders and rounded outer angles, slightly longer than wide, widest at the outer angles. Four arm spines in the free basal joints, cylindrical, tapered, blunt ; the second from above which is the longest, is slightly longer than the corresponding arm joint, while the lewdest one, which is the shortest, is slightly shorter than the same. A short distance out from the disk, there are only three arm spines on either side of each arm joint, the uppermost being ^the longest. A single large, oval, flat tentacle scale to each pore, except the first v>rhich has usually two scales. Colour in alcohol bluish. This new species very much resembles A. sexracUa (Duncan, 1889), but differs from it in the fewer and more widely spaced disk spines, in the quite distinct disk scales and radial shields, in the more numerous oral papilla3, in the longer and narrower oral shields, in the shape of the dorsal and ventral arm plates and in tlie longer arm spines. 182 . AET, 2. — H. MATSUMOTO : Key to Japanese species of Aniphiodia. A — Radial shields fully in contact in pairs ; ventral arm plates pentagonal, as wide as, or wider than, long ; four arm spines ; two tentacle scales crater odometa. AA — Eadial shields divergent, slightly or not at all in contact in pairs ; ventral arm plates quadrangular, longer than wide ; outer two oral papillae pointed ; tentacle scales very rudimentary. a — Interbrachial ventral surfaces finely squamated ; four or five arm spines ; one or two rudimentary tentacle scales euraspis. aa — Interbrachial ventral surfaces covered by a naked skin ; four or five arm spines ; single rudimentary tentacle scale to each pore. .psUocliora. AtnplHoäia craterodo^neta Clakk. AmpModia craierodometa : Clark, Bull, U. S. Nat. Mus., LXXV, 1911 p. 155, fig. 65. Yezo Strait; 533 fathoms (Claek). Okhotsk Sea; 58-69 fathoms (Claek). Off Saglialin ; 21-32 fathoms (Claek). Gulf of Tartaiy ; 318 fathoms (Claek). Off Korea ; 116 fathoms (Claek). Alaska (Claek). Bering Sea (Claek). Arctic Ocean (Claek). Aniphiodia euraspis Clark, Ampln'odia eurasjyis : Clark, Bull U. S. Nat. Mus , LXXV, 1911, p. 158, fig. 67. Sm^uga Gulf; 211-293 fathoms (Claek). Off Korea ; 184-250 fathoms (Claek). Gulf of Tartary;"318 fathoms (Claek). Bering Sea (Claek). California (Claek). Ainphiodia psilochora Clark, Ampldodia p&ihcJiora : Clapj;, Bull. U. S. Nat. Mus., LXXY, 1911, p. 165, fig. 72. MONOGIiArH OF JAPANESE OPHIUROIDEA. 183 Sagami Sea ; 202-622 fathoms (Claek). Off Ose Zaki, Suruga Gulf ; r)5-()') fathoms (Claek)- Off eastern Japan ; 440 fathoms (ClaeiO. (fl>h loph vagin us japoniciis Matsumoto. Ojjhiop/i ragmvs jcrponivvs : jNIatsumoïo, Proc. Acad. Nat, Sei. Philadel- phia, 1915, p. 70. Numerous specimens ; off Namami, Kagoshima Bay ; 8-15 fathoms. Two specimens ; Enoura, Suruga. Numerous specimens ; off Oginohama, Eikuzen ; 10-20 fathoms. Diameter of disk 7 ram. Length of arms 45 mm. Width of arms at base 1 mm. Disk five-lobed, with very convex interbrachial borders, covered with very fine, imbricating scales, ^ix primary plates more or less distinguishal)le at the central region. Iladial shields semilunar, one-third as long as the disk radius, twice as long- as wide, separated ovlIj at the proximal end, oljtusely pointed within. A row of large and squarish marginal scales occurs on the disk. The scales of tlie interbrachial ventral surfaces just outside the marginal scales are turned up, so as to form a sort of fence. The marginal scales are more; elevated than Fig. 48. Op/ilnphraijmvs japonlruf;. a. From above, h. Erom below, c. Side view of three arm joints near disk. 184 AKT. 2. — H. MATSUMOTO : the arms, and the interbrachial region of the disk is strongly- convex below. Genital shts long. Oral shields rhomboidal, with the inner sides much longer than the outer, with acute inner and rounded lateral and outer angles, much longer than wide. Adorai shields triangular, tapered within to a point, and not meeting with each other. Four oral papillœ on either side of the oral angle, close-set, continuous, subequal and blunt, the innermost one somewhat stouter. Dorsal arm plates oval, large, outer border curved, inner border strongly convex, forming part of a circle ; as wide as the arms, twice as wide as long, a little in contact with each other. Lateral arm plates inserted like so many wedges between the successive dorsal plates above and ventral plates below ; well separated above but closer below. First ventral arm plate very small, quadrangular, much wider than long. Those beyond pentagonal, with very large inner angle and slightly notched outer border, wider than long, only a little in contact with each other. Three arm spines, conical, subequal, blunt, nearly as long as the corresponding arm joint. Two tentacle scales, very flat and thin ; abradial one smaller than the adradial, and overlapping its base. Colour in alcohol light yellow. This species somewhat resembles 0. aßnis Duncan, especial- ly in the number of the oral papillae, but differs from it in the shape of the radial and oral shields and of the dorsal arm plates. In my opinion, Ampliipholis andreœUàTKE]^, 1872, Amphiiira jjrœstans Kœh- LEK, 1905, and Amphiodla perlei'cta Clakk, 1911, are referable to OpJiiophragmiis, each showing certain affinities to the present species. I have observed also in younger specimens of the present species that, the marginal disk scales are so prominent as to form a hem-like MONOGRAPH OF JAPANESE OPHIUROIDEA. ]85 row of denticles, so that the disk border is serrate, quite as in Clark's fig. 68d of 0. perierctus. Key to Japanese species of Amphipholis, A — Radial shields perfectly joined in pairs, a — Arms three to four times as long as the disk diameter; dorsal arm plates distinctly shorter than the corresponding arm joint, without any streak along the median line. h — Disk scales thin, rather indistinct from one another, so that the surface of the disk is very smooth ; radial shields not very wide, about two and a half times as long as wide, the united width of each pair being nearly equal to the width of the corresponding arm base ; three arm spines, of which the uppermost one is the longest, the lowest one the shortest, and the middle one the stoutest but not distinctly compressed Japonica. hh — Disk scales thickened along the free margins, concave, very distinct from one another, so that the surface of the disk is not very smooth ; radial shields very wide, about twice as long as wide, the united width of each pair much exceeding the width of the corresponding arm base ; tliree, sometimes four, arm spines, of Avhich the uppermost one is the longest, and the middle one the shortest and stoutest and distinctly compressed sobrina. CM — Ai'm seven to eight times as long as the disk diameter; dorsal arm plates as long as the corresponding arm joint, with a distinct white streak along the median line pugetana. cLl — ßadial shields divergent, being nearly or entirely separated from each other by a row of scales; three arm spines, of which the lowest one is the longest ; adradial tentacle scale much larger than the abradial Icocldi. 186 ART. 2. — H. MATSÜMOTO Awiphipholis Japoniea Matsx .moto. Amphipholis japoniea : Matsumoto, Proc. Acad. ZSat. 8ci. Pliiladelpliia, 1915, p. 71. Numerous specimens ; Aral Beach, ÙMisaki Marine Biological Station. One specimen ; Tomo, Bingo. One specimen ; Asami Bay, Tsushima. One specimen ; Akuné, Satsuma. Several speci- mens ; Shimabara, Hizen. Diameter of disk 3 mm. Length of arms 1 2 mm. Width of arms at base 0.5 mm. Disk circular, covered with fine, thin. im])ricating scales, which are again covered over by a thin skin, so that the surface of the disk is very smooth. Radial shields comparative- ly large, pear- seed shaped, slightly shorter tlum half the disk radius, al>ou!f- two and a lialf times as long as wide, completely joined in pairs, ex- cei>t at the very inner ends, where tlicy are separated from each other by a triangular, wedge- shaped scale. Inter- bra chial ventral sur- facivs covered also Fig. 49. Amph'qiltolis japoniea. x27. u. From above. h. From below, c. Side view of two arm joints neai- disk. MONOGRAril OF JAPANESE OrHIUEOIDEA. 187 with fine, imbricating scales, which aiv only slightly finer than those of the dorsal side. Genital slits long. Oral shields rhomboidal, with very acute inner and rounded lateral and onter angles, inner sides longer than the outer ; longer than wide. Adorai shields triangular, long and narrow, wider outwards than inwards, meeting with each other within. Three oral papilla? on either side ; the inner two are rounded, while the outermost one is very large, wide and operculiform. Teeth quadrangular and stout. Dorsal arm plates triangular, with rounded angles and convex sides, about one and a half times as wide as long, much shorter than the corresponding arm joints, so that the successive plates are separated from one another. Lateral arm plates rather low, distinctly flared outwards, those of the two sides meeting with each other both above and below. First ventral arm plate small, triangular, with convex outer side, wider tlian long. Those l^eyond pentagonal, with acute inner angle, outer lateral angles perfectly rounded, inner sides straight, lateral sides concave, outer side convex ; about as long as, or slightly longer than, wdde, widest at the inner lateral angles. Three arm spines, conical, acute ; the uppermost one is the longest, sliglitly shorter than, or hardly as long as, the corresponding arm joint ; the middle one is the stoutest, but not distinctly compressed ; the lowest one is the shortest. Two small, oval, flat, leaf-like tentacle scales, of which the adradial one is slightly larger than the abradial. Colour in alcohol white or light yellow. In life, the disk is light reddish brown. I have compared these specimens with those of A. sqi/ainala (Delle Chiaje) from Naples, and found only two trifling differences between them. In the Japanese specimens, the arms are three to 188 ART. 2. — H. MATSUMOTO : four times as long as the disk diameter and the ventral arm plates have a convex outer border and rounded outer lateral angles, while in the Neapolitan specimens, the arms are two and a half to three times as long as the disk diameter and the ventral arm plates have straight outer border and not rounded outer lateral angles. The difference in the arm length may not be very important in the question of the distinctness of the two species, because in A. sqiiamata the arms are said to vary from two and a half to four and a half times as long as the disk diameter. I there- fore look upon the difference in shape of the ventral arm plates as the only distinctive character of the present species as compared with A. squamata. In the shape of the ventral arm plates, the present species much resembles A. australiana Clakk, but differs from it in the more numerous disk scales of the dorsal side and coarser disk scales of the ventral side. The radial shields have each a white spot at the outer end, as in A. squamata. The present species is common in the neighbourhood of Misaki and is found living under stones in fine sand. As to its sensitive- ness to the coarseness of the sand, I made the following observa- tions at Aral Beach. In the summer of 1910, there were at first among the rocks numerous spots covered with fine sand, and this species was found very abundantly ; but after a tempest, it was reduced to small numbers, owing to the fact that the spots with fine sand became very scanty. In the summer of 1911, spots with fine sand were very scanty, and this species was also few. In the summer of 1912, the beach was entirely covered with coarse sand, and the species could not be found any longer. This ophiuran is very quick in motion, and can instantly conceal itself in the sand, when the stone is turned up. The present species is undoubtedly viviparous like ^l. squamata. MONOGEAPH OF JAPANESE OPHIUKOIDEA. 189 Til summer, the larger individuals contain several embryos. I have once dissected out six embryos from a single adult. Animals containing full grown embryos appear to give birth to them the night after they are placed in an aquarium. Amphipholis sobrhia, sp. nov. Numerous specimens ; Mera-out-Oisegaké, Sagami 8ea ; 300 fathoms. Five specimens ; Sengendzuka-Aoyamadashi, Sagami Sea ; 85 fathoms. Three specimens ; oft' Oshima, Sagami Sea ; 75-85 fathoms. Diameter of disk 3 mm. Length of arms 10 mm. Width of arms at base 0.6 mm. Disk subpentagonal, covered with imbricating scales, which are rather fine, thickened along the free margins, concave, very distinct fi'om one another and coarser inwards than outwards. Six primaries rather distinct, larger than the other scales, separated from one another. Radial shields large, pear- seed shaped, shghtly shorter than half the disk radius, about twice as long as wide, completely joined in pairs, except at the very inner ends, where they are separated from each other by a triangular, wedge-shaped scale ; the united width of each Fig. SO. AmpldphoUs sohrina. x 20. a. From above, h. From below, c. Later.-il view of three arm joints near disk. 190 AKT. 2. — II. MATSUMOTO : pair distinctly exceeds the width of the corresponding arm base. Interbrachial ventral surfaces covered also with imbricating scales, which are much finer than those of the dorsal side. Genital slits long. Oral sliields rhomboidal, with acute inner and rounded lateral and outer angles, inner sides longer than the outer ; slightly longer than, or nearly as long as, wide. Adorai sliields triangular, long, tapered inwards, meeting with each other within. Three oral papilla3 on either side, the inner two are I'ounded, and the outer- most one is very large, wide and operculiform. Teeth quad- rangular and stout. Dorsal arm plates triangular, with rounded outer lateral angles and convex outer side, one and a half times as wide as long, shorter than the corresponding arm joint, so that the succès- sive plates are not in contact with each other. Lateral arm plates low, flared outwards, those of the two sides slightly meeting with each other both above and below. First ventral arm plate exceedingly sm.all, pentagonal, longer than wide. Those beyond pentagonal, with rounded outer lateral angles and convex outer side, wider than long, widest at the inner lateral angles, not in contact with each other, except the first two or three plates. Three, or sometimes four, arm spines, of which the uppermost is usually the longest, and tlie middle one is the shortest and stoutest and is distinctly compressed at the base. Two very small, oval, flat, leaf-like tentacle scales, of which the adradial one is slightly larger than the al^radial. Colour in alcohol : wählte or yellow, except the dorsal side of the disk, which is light bluish ; radial shields with a white spot at the outer end, as in A. squamaia and A. japonica. The present species is very near to A. japonica, but diiEfers MOXOGEAPH OF JAPANESE OPHIUEOIDEA. 191 from it ill the coarser and more distinct disk scales, in the much larger radial shields, in the wider ventral arm plates and in the middle arm spines b^niig the shortest and distinctly compressed at the base. The specimens from Mera-out-Oisegaké have stouter arms than those fi'om Sengendzuka-Aoyamadashi and from off Oshima. This difference may probably be due to bathymetrical factor. I regard the specimens from Mera-out-Oisegaké as the typo of the present species. Aiiipfiipholis pugetmia (Lyman). Ampidura pmjetana : Lyman, Proc. Boston See. Nat. Hist., YII, 1860, p. 193^^; Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, p. 125; Lyman, Eep. Clialleuger, V, 1882, p. 145; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, p. 305. AmpMpholis pngetana : Vereill, Trans. Conn. Acad., X, 1899, y>. 312 ; McClendon, Univ. California Publ. Zool., VI, 1909, p. 43, PI. II, figs. 12 «t 13 ; Claek, Bull. Ü. 8. Nat. Mus., LXXV, 1911, p. 166, figs. 73. One specimen (belonging to Mr. E. Saitô, Imperial Museum at Uyeno, Tokyo); Aomori Bay. Four specimens (belonging to Mr. Sasaki, Coll. Agricult., Sapporo); Oshoro, near Otarunai, Hokkaido. One specimen (belonging to Mr. H. Asano, Imperial Bureau of Fishery); off Kitami ; 60 fathoms. Pnget Sound. California. Washington. Alaska. The largest one of these specimens measures 4 mm. across the disk and 28 mm. in the arm length. In the larger specimens, the disk scales are very fine and exceedingly numerous, and the primary plates are indistinguishable ; while in the smaller speci- 1) This paper wn« not seen by me. 192 APtT. 2. — H. MATSUMOTO mens, the disk scales are com- paratively coarse and the primary plates very distinct. The arms are long and very slender, being widest at about one-fom^th the entire length from the base. The dorsal arm plates are nearly as wide as the corresponding arm joints, and about one and a half times as wide as long. The ventral arm plates are slightly wider than long and have a notched outer border. The dorsal, as well as the ventral, arm plates are not in direct contact with one another, being slightly separated by wedge-like prolongations of the lateral arm plates, except the first two ventral arm plates, which are joined to each other. The colour in alcohol well agrees with Lyman's description, but the arms are irregularly banded with grayish shade. I am informed by Mr. Sasaki that the animals are found attached to the rhizomes of Zostera marina. It is interesting to find this species occurring in the northern waters of Japan, since it has been known only from the eastern coast of the North Pacific. Fig. BI. AmpMjpholis jmgetana. xl2. a. From aboYe. h. From below. c. Side view of three arm joints near disk. Amphipholis Ji'ochii Lütken. Ampldjyholis kochii : LÜTKEN, Oph. Nov., 1872, p. 10, PI. I & 11, fig. 6. Amphiura JcocJiii : Lyman, Kep. Clialleriger, V, 1882, p. 146. Numerous specimens ; Aral Beach, Misaki Biological Station. MONOGEArn OF JAPANESE OrHIUROlDE.A . 193 Numerous specimens ; Misaki. One specimen (belonging to Mr. E., Saitô, Imperial Museum at Uyeno, Tokyo) ; Aomori Bay. Uladiwostock (Lütken). This species is very variable. The disk scales are often coarse and very irregular. The primary plates are sometimes visible. The oral shields are rhomboidal, with the inner sides longer than the outer, and longer than wide ; so that in Lütken's fig. 6b, the upper right one is most natural. The ratio of the arm length to the disk diameter is greater in smaller specimens ; Fig. 52. Amphipholift koclm. a. From above. x8. h. From below. x8. c. Side view of two arm joints near disk x 8. (?. From above. x6. e. From below. x6. /. Side %iew of three arm joints near disk. x6. SO that the size contrast of the arms is not so striking between two specimens, one of which is 4 mm. and the other is 7 mm. across the disk. In the smaller specimens the disk is flat, sub- 194 ART. 2, — H. MATâUMOTO : pentagonal, with nearly straight interbrachial borders ; while in the larger ones the disk is swollen and the interbrachial regions bulge out more markedly. The radial shields do not take part in this .bulging process, and are consequently deeply sunken, while their distance from the centre of the disk is not much greater than in the smaller specimens. The radial shields are distinctly separated in the smaller specimens, but more or less meet with each other in the larger ones. The colour in alcohol, as in life, is also variable. Some specimens are gray or yellowish gray, and spotted on the arms with dark gray ; others are dull yellow, with some disk scales gray ; the radial shields are usually light yellow, hemmed with dark gray. Specimens with coarse and irregular disk scales are darker in colour. Amphiura (Foebes, 1842) Verrill, 1899. Syn. OpMonephthjs Lütken, 1869. I look upon Ophionephthys Lütken as a synonym of Amphiura s. str. The reduction of the disk scales which is a characteristic of Opliioneplithys is also observed in Amphiura acrystata, as well as in some Amphiura formerly referred to Hemilepis, while the absence of the tentacle scales is characteristic of certain Amphiura formerly referred to Ophiopeltis. There is not a single character by which Ophionephthys can be satisfactorily distinguished from Amphiura. Key to Japanese species of AmjiMura. A — Tentacle scales present. a — Two tentacle scales to each pore. MONOGRAPH OF JAPANESE OrHIUKOIDEA. I95 h-Two, or sometimes three, distal oral papiUœ on either side, besides an additional one, which lies deep within the oral slit. c— Three arm spines, subeqnal, or the middle one slightly longer than the other two ; disk distinctly squamated both above and below, with very prominent primaries ; radial shields large, rather wide, each pair being separated from each other by a row of usually three large scales; ventral arm plates wider than long Jcorcœ. ce— Four arm spines, of which the lowest one is the longest and the third from below the shortest; disk distinctly squamated above, but co^'ered with fine granule-like scales below ; the primaries are indistinct in larger specimens ; radial shields long and very narrow, slightly divergent; outer end of each genital plate, just distal to the radial shield, bearing a projection, which is thorny at the tip ; ventral arm plates longer than wide digitula. hh-^mg\e distal oral papilla on either side, besides an additional one, which lies more or less deep within the oral slit. d — Four arm spines. e— Disk squamated both above and below. /—Disk scales very fine and thin ; radial shields long and narrow, about thrice as long as wide, separated from each other ; adorai shields not meeting with each other; arm spines subeqaul . . .hellis. #— Disk scales rather coarse and thick, some of them projecting like a hemisphere, so that the surface of the disk is very rough; radial shields not very long, about twice as long as wide, separated from each other; adorai shields meeting with each other in the inter- radial line ; the uppermost and the lowest arm spines distinctly longer than the middle two trachydwca. €e-Disk squamated above, but naked below ; radial shields partially joined in pairs ; adorai shields meeting with each other in the interradial line; middle two of the four arm spines stouter and blunter than the uppermost and lowest microdiscus. dd—Fixe to seven arm spines; disk naked at least below. 196 ART. 2. — H. MATSUMOTO : g — Disk rather thick, distinctly squamated above ; radial shields short ; lower arm spines spur-shaped and rough at the tip euopla. gg — Disk very thin, mostly naked, the scales persisting only around the radial shields, which are long and rather narrow ; arm spines conical, not rough acryskda. aa — Single tentacle scale to each pore. Ji — Three arm spines. '/ — Disk distinctly squamated both above and below. ,/ — Disk scales rather fine ; radial shields very wide, widest near the inner end ; oral shields rounded, longer than wide ; adorai shields Avidely separated from each other in the interradial line ; dorsal arm plates separated from one another by the lateral arm plates ; ventral arm plates pentagonal, longer than wide, only slightly in contact with one another ; arm spines subequal acacia. jj — Disk scales coarse ; radial shields not very wide, wider outwards oral shields oval, much wider than long ; adorai shields widely meeting with each other in the interradial line ; dorsal arm plates widely in contact with one another; ventral arm plates hexagonal or heptagonal, widely in contact with one another ; middle arm spine longer than the other two pycnodoma. a — Interbrachial ventral surfaces naked ; radial shields long and narrow ; oral shields rounded pentagonal, wider than long ; adorai shields not meeting with each other in the interradial line ; distal oral ]3apilla long and spiniform ; dorsal, as well as ventral, arm plates slightly in contact with one another ; arm spines subequal . . carcliara. hh — Four to seven arm spines. k — Four or five arm spines, of which the lowest is the longest and the third from below the shortest. I — Four, rarely five, conical arm spines, of which the maximum length does not exceed twice the corresponding arm joint ; disk scales ill- defined, not distinctly imbricated, but pavement-like ; radial shields very long and narrow, bar-like ; oral shields large, much wider than ÎIONOGEAI'H OF JAPANESE OPHIUKOIDEA. 197 long ; distal oral papilla not much larger tlian the apical papilla and tentacle scales, rather thick ; dorsal arm plates very convex in longitudinal secticm ; ventral arm plates not keeled iris. II — Five flattened arm spines, of which tlie maximum length is about thrice the corresponding arin joint ; disk scales distinctly imbricated ; radial shields not xevy long, pear-seed shaped ; oral shields small, nearly as long as wide ; distal oral papilla much larger than the apical papilla and tentacle scales, very flat and thin ; dorsal arm plates rather straight in longitudinal section ; ventral arm plates with a median keel iridoiths. Jck — Six or seven arm spines. m— Six arm spines, of which the third or fourth from above is the longest; radial shields long and narrow; oral shields diamond - shaped, about as long as wide ; distal oral papilla subspinous, long, compressed li'dkeui. mm— Six or seven arm spines, of which the lowest one is the longest; radial shields extremely small, short and narrow ; oral shields oval, about as long as wide; distal papilla very large, fan-shaped, flat., micraspis. AA — Tentacle scales absent ; disk mostly naked, the disk scales persist- ing only around the radial shields ; numerous arm spines, of which the lower ones are rough at the tip. n — rive to seven arm spines ; dorsal and ventral arm plates wider than long. o — Five arm spines near disk, but four more distally ; dorsal arm plates very large and wide even at the arm bases; arms twelve or thirteen times as long as the disk diameter œstuarii. 00— Six or seven arm spines near disk, but five or six more distally; dorsal arm plates very small and rudimentary at the arm bases ; arms exceedingly long, more than thirty times as long as the disk diamete]- vadicdla. ?2?i — Ten arm spines; dorsal and ventral arm plates much less wide than 198 AET. 2. — H. MATSUMOTO long; arms about fourteen times as loug as the disk diameter ... . ecnooniotata . Ainphûira koreœ Duncan. AmpMura Jcorece : Duncan, Journ. Lirm. Soc. Loudon, XIV, 1879, ]). 166. PI. X, figs. 18 & 19. AmiMura corece : Lyman, Kep. Challeuger, Y, 1882, p. 146. AmpJiipholis corece : Veeeill, Transact. Connecticut Acad., X, 1899, p. 321. Amphiura diomedece : Lütkfn & Moetensen, Mem. Mus. Comp. Zool., XXn, 1899, p. 151, PI. Xni, figs. 1-7; Kœhlee, Exp. Siboga, XLY, Pt. 1, 1904, p. 86; Claek, Buli, U.S. Nat. Mus., LXXV, 1911, p. 140. A Numerous specimens ; Ofcaba, Sagami Sea ; 500 fathoms. Off Manazuru Zald, Sagami ; 120-749 fathoms (Claek). Off Port Arari; 400-726 fathoms (Claek). Sumga Gulf; 140-503 fathoms. Off Om.ai Zaki ; 475-662 fathomxs (Claek). Off eastern Japan ; 39 fathoms (Claek). Sea of Japan ; 79 fathoms (Claek). Kagoshima Gulf ; 58 fathoms (Claek). Pacific side of Central America. Malaysian waters (Kœhlee). That the specimens at hand are referable to A. diomedece appears to me certain. But an examination of Duncan's description and figures of A. Jwrece, which is considered by several authors to belong to what is now called Amphipholis, has con- vinced me that Duncan's species in question is merely an Amphiura s. Fia. S3. AmpJnura koreœ. x7. a. ^ . i t i • i i • i _ 'x.\ From belo.-. ?>. Ventral view of an Str., aud IS probably idcutlCal With oral angle, c. Ditto. ^ dlOmedeW. DUNCAN did UOt MONOGIL\PH or JAPANESE OPHIUEOIDEA. 19 mention the size of his type specimen ; but judging from the magnification of his figures, his specimen must have been much smaller than Lütken à Moetensen's type. The smaller specimens in my hands appear to me to correspond well to Duncan's descrip- tion and figures, the only essential difference being the shape of the oral shield, which is lyre-shaped and very wide outwards in Duncan's figure. But the oral shields, adorai shields and oral papilloe, &c. of A, dlomedece are subject to great variation, as stated by Lütken & Mortensen. Therefore, I am inclined to look upon A. korece and A. dlomedece to be conspecific, the former having the priority. In the specimens at hand, the arms are mostly four to six times as long as the disk diameter, so that they are usually longer than in Duncan's and Kœhler's specimens, but usually shorter than in Lütken & Mortensen's. In some of the oral angles, there may be one to three additional infradental papilla) besides the ordinary pair. The disk is usually light bluish and the arms yellowish or light brownish in colour in alcohol. Amphiura digitula (Claek). AmpUodia digitula : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 162, fig. 70. Four specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Off Ose Zaki, Suruga Gulf; 45-48 fathoms (Clark). Sea of Japan ; 70 fathoms (Clark). In my opinion, the present species is evidently referable to Amphiura s. str., though Clark refers it to Amphiodia. The distal oral papillse are usually two in number on either side, both arising from the adorai shield. In some cases, there are three 200 AET. 2, — H. MATSÜMOTO Pig. B4. Amphiura digüula. x 15. a. From above, h. From below, c. Ventral view of an oral angle, d. Side view of three arm joints near disk. distal oral papillae, of which one arises from the oral plate (fig. c); besides, there is an additional papilla, which evidently cor- responds to the scale of the first oral tentacle, just inside and above the inner- most one of the distal papillae, thus almost approaching the type of the oral papillae of Ainphioplus. The disk is light bluish and the arms whitish in colour m alcohol. Ainx>Jilura hellis Lyman. AmpMura hellis : Lyman, Ball. Mus. Comp. Zool., VI, Pt. 2, 1879, p. 19 ; Lyman, Rep. Challenger, V, 1882, p. 127, PI. XVni, figs. 4-6, PI. XL, figs. 16-18 ; Clark, BuU. U. S. Nat. Mus., LXXV, 1911, p. 140. Sagami Sea ; 345-775 fathoms (Lyman). Off Manazuru Zaki, Sagami Sea ; 120-265 fathoms (Claek). Off Osé Zaki, Suruga Gulf; 63-75 fathoms (Claek). Off Omai Zaki; 624 fathoms (Clark). Off Kii ; 649 fathoms (Claek). Lat. 19° 10' S., long. 178° lO'E.; 210-610 fathoms (Lyman). MONOGRAPH OF JAPANESE OPHIÜROIDEA. 201 Atnphiura trachydisca Clark. AmpMura trachydisca : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 149, fig. 60. One specimen ; Yahagi-gaké, Sagami Sea ; 310 fathoms. Two specimens ; Nakano-yodomi, Sagami Sea ; 210 fathoms. Numerous specimens; Okinosé, Sagami Sea; 100 fathoms. Numerous specimens ; Sagami Sea. Off Heda; 161-167 fathoms (Clark). Suruga Gulf; 108-148 fathoms (Claek). Uraga Channel ; 88 fathoms (Clark). The largest one of these specimens is 15 mm. across the disk. Those from Okinosé are somewhat different from the other speci- mens, which agree quite well with Clark's type, and have an almost smooth disk and relatively narrower arms, though there is no doubt that they belong to this species. Amj^hiura microdiscus Duncan. Hemipholis microdiscus : Duncan, Journ. Linn. Soc. London, XIV, 1879, p. 467, PL X, figs. 20-22 ; Lyman, Rep. Challenger, V, 1882, p. 158. Korea Strait ; 51 fathoms (Duncan). According to Duncan, this species has two pairs of oral papillaß to each oral angle, the apical ones being infradental and the distal ones arising from the adorai shields. There is therefore no doubt that, this species is a genuine Amphiura and not a Hemipholis. Amphiura eiiopla Clark. Ampliiura Candida : Marktanner-Turneretscher, Ami. K. K. Naturh. Hofmus, II, 1887, p. 299. (Non Ltungman, 1866.) 202 ART, 2. — H. MATSUMOTO AmpJiiura enopla : Claek, Bull. U. S. Nat. 3Ius., LXXV, 1911, p. 144, fig. 57. One specimen ; Enoshima, Sagami. Numerous specimens ; Misaki. Numerous specimens ; Izubara, Tsushima. Off Osé Zaki, Suruga Gulf ; 45-60 fathoms (Clakk). The specimen from Eno- shima is the largest of all, being 6 mm. in the disk diameter, and is the only one in which the primaries are invisible. The others are all small, and have conspicuous primaries and dis- tinctly squamated interbrachial ventral surfaces. The distal oral papilla is not so long and spini- form as in Claek's type, and the oral shields are not so long, but slightly longer than wide. Further, in the smaller specimens, they are as wide as, or wider than, long. The arm spines are much flattened, except in the basal region of the arms, and are thorny at the end, with the exception of the uppermost one or two. In the smaller specimens, they are less flatten- ed and less thorny. Fig. SS. Amphiura euopla. a. From above. X 8. h. From below, x 8. c. Young speci- men, from above, x 14. d. Side view of two arm joints near disk, x 8. e. Dorsal view of three arm joints near disk, x 8. /. Arm spines of one side of an arm joint near disk. xli. MONOGRAPH OF JAPANESE OPHIÜEOlüEA. 203 Amplnura aerystata Claek. AmpJinira aerystata : Clapjv, Bull. U. S. Nat. Mus., LXXV, 1911, p. 145, fig. 58. Numerous specimens ; Okinosé (a submarine bank), Sagami Sea ; 85 fathoms. Off Tsurugi Zaki ; 110-259 fathoms (Claiik). Off Kii ; 191 fathoms (Claek). Sea of Japan; 59 fathoms (Clapjv). Southwards from Hokkaido ; 269-4G4 fathoms (Clark). California ; 8-33 fathoms (Claek). In the larger specimens, the disk coverings remind us of those of the Ophionephthys-tyYjQ, the tentacle scales being however well developed. This species may be brought forward to prove the small taxonomic value of the reduction of the disk scales, which is the character separating Ophicnephthys from Amphiura. Atnpthiura acacia Lyblyn. AmpMura acacia : Lymax, Bull. Mus. Comp. Zool., VI, Ft. 2, 1879, p. 21, PI. XI, figs. -292-294 ; Lyman, Rep. Challenger, V, 1882, p 130, PL XVI, figs. 15-17. Yenshû Sea ; 565 fathoms (Lyman). AmpJiiura pycnostonia Clark. AmjMura pycnostoma : Claex, Bull. U. S. Nat. Mus., LXXV, 1911, p. 151, figs. 62. Off Kii, Kumano Sea ; 440 fathoms (Claek). Ampjliiura carcliara Clark. Amphiura carcliara : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 142, fig. 55. Sagami Sea; 614 fathoms (Claek). 204 ART. 2. — H. MATSÜMOTO : Bering Sea. Alaska. Washington. AtnpJiiura iris Lyman. AmpJdura iris : Lyrian, Bull. Mus. Comp. Zool., VI, Pt. 2, 1879, p. 23, PI. XI, figs. 302-304; Lyman, Hep. Clialleuger, V, 1882, p. 132, PL XVI, fig. 4-G; Kœhlee, Exp. Siboga, XLV, Pt. 1, 1904, p. 81. One specimen ; Mera-out-Oisegaké ; Sagami Sea ; 300 fathoms. Seven specimens; Okinosé (a submarine bank), Sagami Sea; 85 fathoms. Sagami Sea; 420-775 fathoms (Lyman). Malaysian waters; 1,595m. (Kœhlee). The first specimen, which is the largest, measures 5.5 mm. across the disk and 25 mm. in the arm length. The disk scales are fine, ill-defined, convex and pavement-like, so that the dorsal surface of the disk appears rough. The radial shields are long, narrow, bar-like, about half as long as the disk radius, about four times as long as wide, completely separated from each other, those of a pair almost parallel. The distal oral papilla is sub- quadrangular, with truncated or indented free end. The apical one is also squarish, but slightly smaller than the distal one. The additional papilla, which corresponds to the scale of the first oral tentacle pore, is also squarish, with truncated or indented free end, smaller than the distal, but larger than the apical one. The dorsal arm plates are rather oval and slightly in contact witli one another ; the surface is convex both transversely and longi- tudinally, so that the plate is more or less humped. The ventral arm plates are subpentagonal, slightly in contact with one another. The arm spines are variable in length ; the ratio of their length to the corresponding arm joint in different parts of an arm may be shown approximately as- follows : — MONOGKArH OF JAPANESE OPHIUROIDEA. 205 "~- Number of lateral "^■^ — -^^^^ arm plates. Number of ""—^-.^^^ arm spines. --—-_ 1st. 5th. 10th. 15th. 20th. 30th. Uppermost 4th. from below 3rd. from beloAV Süd. from below . . Lowest 0.5 1.2 1.6 1.3 0.9 ].l 1.4 1.2 0.9 1.1 1.8 l.'i 1.0 1.2 1.7 1.2 1.1 1.1 1.6 1.2 1.1 1.4 The tentacle scale is oval and leaf-like, but relatively smaller than in the next species. Ainplihira Irldoides, sp. nov. -AmpJmira cjlahra : Maektanner-Tl'RNERETSCHER, Ann. K. K. Naturh. Hofmus., II, 1887, p. 300. (Non Lyman, 1879.) Twelve specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Diameter of disk 4 mm. Length of arms 20 mm. . Width of arms at base 0.9 mm. Disk subpentagonal, with more or less straight interbrachial borders, covered with rather fine, imbricating scales, which are finer outwards than inwards. Six primaries prominent, distinctly larger than the other scales. Radial sMelds rather large, pear- seed-shaped, about two-fifths as long as the disk radius, two and a half to three times as long as wide, widest at about one -third of the entire length from the outer end, entirely separated from each other, divergent. Interbrachial ventral surfaces covered also with fine, imbricating scales, which are finer than those of the dorsal side. Genital slits long, extending from the oral shields to the disk border. Oral shields small, rhomboidal, with obtuse inner and rounded 206 AP.T. 2.— H. MATSUMOTO : lateral and outer angles, about as wide as long. Adorai shields also small, triangular, with con- cave adradial side, longer than wide, tapering inwards, nearly meeting with each other in the interradial line. Oral plates rather short, very narrow. There is a small, naked, de- pressed space between the adorai shields and the oral plates of each oral angle. Two oral papillae on either side ; the distal one, arising from the adorai shield, is very large, oval, fiat, thin and leaf-hke ; the apical one is infradental, small, short, very thick. Teeth quadrangular and thick. Arms stout in comparison to the disk diameter, more or less cylindrical, wider than high, very gradually tapering outwards. Dorsal arm plates fan-shaped, with convex outer border, wider than long, wider outwards than inwards ; the surface is nearly level longitudinally, but decidedly convex transversely. Lateral arm plates rather high, those of the two sides not meeting with each other above or below, except in the very distal arm joints ; they are almost covered over by the arm spines. First ventral arm plates not very small, pentagonal, wider than long. Those beyond hexagonal, with the shortest inner lateral sides and longest and convex outer border, lateral sides concave, corresponding to the tentacle pore ; about as long as, or longer than, wide, widest at the outer ends of the tentacle pores, faintly Fig. S6. Amphiura iridoides. x 14. a. From above. l>. From below, c. Side view of three arm joints near disk. MONOGRAPH OF JAPANESE OPHIUROIDEA. 207 keeled along the longitudinal median line, widely in contact with each other. Five arm spines, long, flattened, acute, unequal in length ; the ratio of their length to the corresponding arm joint in different parts of an arm may be shown approximately as follows. ~ - ^^^ Number of lateral ~~~~--..,.,^^^ arm plates. Xiimber of ~"~~-^^_^ 1st. 5th. 10th. 15th. 25th. 35th. arm spines. Uppermost 1.4 1.4 1.3 4th. from below ].2 1.2 1.1 1.1 1.1 3rd. from below 1.0 1.0 1.0 1.0 0.9 2nd. from below 1.0 1.3 1.4 1.3 1.2 1.1 Lowest 1.3 1.5 2.0 3.0 1.8 1.4 Thus, the lowest spine is always the longest and the third from below the shortest ; the lowest one is longest at a certain distance from the disk, being there about thrice as long as the corresponding arm joint. Single, very large, oval, flat, leaf-like tentacle scale to each pore. Colour in alcohol : whitish, except the dorsal side of the disk, which is frequently light blue. This new species is very near to A. iris, which occurs also in the Sagami Sea, but differs from it in the distinctive characters already mentioned. Further, A. iridoides appears to me to differ from A. glabra Lyman in the very prominent primary plates, in the distinctly squamated interbrachial ventral surfaces, in the much smaller and narrower oral shields, in the not very distinctly separated adorai shields, in the larger, oval, flattened distal oral papillae, in the longer and flattened arm spines, and in the larger tentacle scales. Judging from Marktanner-Turneeetscher's description, the specimen which he referrs to A. glabra appears to me likely to belong to A. iridoides. 208 AKT. 2. — H. MATSUMOÏO : Aniphiuva lütkeni Duncan AmpMura liUkerà : Duncan, Joiirn. Linu. Soc. London, XIV, 1879, p. 464, PI. X, fig. 17. (Non Lyman, 1882.) AmpJiiura duncani : Lyman, Rep. Challenger, V, 1882, p. 143 ; Loriol, Rev. Suiss. Zool, I, .1893, p. 403 ; Kœhlee, Res. Exp. Siboga, XLV, Pfc. 2, 1905, p. 33. Korean Seas (Duncan). Malaysian waters (Kœhlee). The specific name of this species was changed by Lyman, because he thought that, after the union of Ljungman's Amphi- pholis with Amphiura, Amphiura I'ldkeni was preoccupied by Amphipholis lütkeni of Ljungman. Now, the latter species has been referred by Veeeill to Ophiocnida or possibly to Amphiodia, so that the specific name of the present species can be revived. Amphiwra /inicraspis Claek. AmpMura micraspls : Claek, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 150, fig. 61. Off Omai Zaki ; 36 fathoms (Claek). Ainphiura œstuarii Matsumoto. Ampldura œstuarii : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 73. Numerous specimens ; Aburatsubo, Misaki Marine Biological Station. Diameter of disk 6 mm. Length of arms 75 mm. Width of arms at base 0.8 mm. Disk five-lobed, covered by a soft naked skin, except along MÖNiOCxRAPH OF JAPANESE OPHIUROIDEA. 209 the abradial and inner borders of the radial shields, where it is covered by fine, imbricating scales, arranged in four or five rows on the inner border but in only one on the outer part of the abradial side, and finer in- wards than outwards. Naked part of radial shields large, pear-seed-shaped, more than half as long as the disk radius, more than twice as long as wide, hardly in contact with- out, slightly divergent within. Interbrachial ventral surfaces also covered by a soft naked skin. Genital slits long. Oral shields rhomboidal, or pentagonal with a very short inner side and a much round- ed outer angle ; wider than long ; madreporic shield much larger and almost circular. Fig. 57. Amphmra œsiuar'd xlO. a. From above, h. From below, c. Dorsal view of three arm joints at about one-third of the arm length from the base. d. Side view of three arm joints near disk. e. Side view of three arm joints at about one-third of the arm length from the base. Adorai shields triangular with concave adradial side, tapered within, where they do not meet. Two pairs of oral papilla? in each oral angle ; apical ones oval and very stout ; distal ones spiniform, obtuse, arising from the adorai shields. Besides, there occurs one spiniform papilla on either side just above the distal one. Dorsal arm plates transversely oval, twice as wide as long, successive plates in contact. Lateral arm plates not very prominent, almost concealed by the arm spines, not meeting above 210 ART. 2. — H. MATSUMOT.): or below, not in contact on the sides, bnt separated by a naked space. First ventral arm plate very small, pentagonal or quadrangular, wider within than without. Those beyond quad- rangular, with convex inner side, notched outer side and rounded outer angles ; wider than long, except the basal one or two ; the successive plates are not in contact but separated by narrow spaces, where the ventral ends of the lateral arm plates are wedged in. Arm spines five in the basal joints, but four in the middle part of the arms, subequal, but slightly longer downwards, nearly equal to, or a little longer than, the corresponding arm joint ; they are conical and obtuse in the proximal joints, but are flattened more distally ; the second spine from below is especially flattened and rather spur-shaped, with numerous thorns on the very much flattened end ; the lowest one, as well as the third from above is also somewhat thorny at the end. Tentacle pores large, without scales. Colour in alcohol: disk gray, radial shields and arms straw- yellow. A. œstuarii differs from A. phalerata (Lyman, 1874) and the next species, A. radicola, in the shape of the radial shields, in the much shorter arms and in the dorsal arm plates, which are very wide even in the basal arm joints. A. œstuarii appears to be closer to A. phalerata than to A. radicola in the arm spines, which are four or five in number and less flattened and rather conical in the basal arm joints. Further, A. œdnarii diflers from A. ecnomiotata in the shape of the radial shields, and of the dorsal and ventral arm plates, as well as in the fewer and less thorny arm spines. A. œstuarii can be easily obtained together with A. euopla by dredging in the muddy bottom of Aburatsubo Cove. It probably MONOGRAPH OF JArANESE orHIUROIDEA. 211 lives buried in mud, as A. vacUcola does in sand, and I believe that, the reduced disk scales and the numerous and thorny arm spines are correlated with its mode of life. A'tnphiura vadicola Matsumoto.'^ OphionepJd/ujs phalercda : Maektanner-Tueneeetscher, Aun. K. K. Naturhisi Hofmiis., II, 1887, p. 301. (Non Lyman, 1874.) Ampliiura vadicola : Matsumoto, Proc. Acad. Nat. »Sei. Philadelphia, 1915, p. 71. Numerous specimens ; Sakura-jima, Kagoshima Gulf. One specimen ; Chintô, Korea. Diameter of disk 8 mm. Length of arms 260 mm. Width of arms at base 1 mm., at the widest part 1.3 mm. Disk five-lobed, with indented interbrachial borders, covered by a soft naked skin, except along the abradial and inner borders of the radial shields, where it is covered by several rows of fine, imbricating scales. Kadial shields large, long, pear- seed -shaped ; naked part two-thirds to one-half as Fig. 58. Amphiura vadicola. x7. (/. From above. b. From below, c. Dorsal view of three arm joints at about the widest part. d. Ventral view of three arm joints at about the w^idest part. e. Side view of three arm joints somewhat near disk. 1) As to the interesting life mode of this ophiuran, see Prof. Mitsukuri & H.i.EA's " The Ophiurian Shoal " in Annot. Zool. Japon., I, 1897, p. 68. 212 ART. 2. — H. MATSÜMOTO : long as the disk radius, and about thrice as long as wide. Inter- brachial ventral surfaces covered also by a soft naked skin. Genital slits long, (lenital scales not very conspicuous unless the specimen is dried, arranged in a row and overlapping one another. Oral shields pentagonal, witli rounded angles, outer sides longest, inner side a litt'e concave ; madreporic shields* mucli larger and almost circular. Adorai shields triangular, with concave adradial side, neither meeting radially nor intcrradially. There is a more or less conspicuous depression between the oral plates, as in OpJiiothrlx. Two oral papilla.' on either side of the oral, angle, conical, blunt, very stout; the distal one arises from the adorai shield, and is longer than the apical one. Arms exceedingly long, more tlxan tliirty times as long as the disk diameter ; they are widest at about one-third of the entire length of the arms from the base. Dorsal arm plates almost oval, bounded within by two nearly straight lines forming a very large and obtuse angle, and witliout by a curve, which is nearly flat medially but very convex laterally ; about twice as wide as long, successive plates slightly in contact. In the basal joints, they are very small and separated by spaces which are covered by a soft naked skin. T/ateral arm plates not very prominent, almost covered over by the arm spines, not meeting al)Ove or below, nor in contact on the sides but separated by a naked space. First ventral arm plate very small, quadrangular, wider than long. Those beyond quadrangular, wider than long, except the basal one or two, which are as long as, or longer than, wide. They in- crease in size, especially in width, outwards, and become pentagonal beyond the disk, with large and obtuse inner angle, rounded outer angles, and notched outer side ; the successive plates are separated MONOGllAril OF JAPANl«]«: OrniUROlDKA. 213 by very narrow spaces, whero the ventral ends of the lateral arm plates arc wedged in. Tlie ventral arm ])lates are often divided into halves along the median line. Arm spines six or seven near the base, live or six in the middle part of the ai'ms, peg-like, flattened, binnt, longer downwards, nearly equal to, or slightly longer than, tlie corresponding arm joint ; much flattened and thorny at the end, except the uppermost one or two ; the second one from below is spur- shaped and very thorny. Tjarge tentacle pores, unprotected. Colour in alcohol brown; the scales around the radial shields are lighter ; arms grayish brown to gray in the outer parts. Tliis species is very near to A. phdlerala (IjYMan, 1874) but differs from it in tlu^ much larger radial shields, in the not oval but pentagonal oral shields, in the adorai shi(3lds being joined neither radially nor interradially, in the more projecting oral angles, in the dorsal arm plates being in contact witli one another, in the ventral arm })lates being separated from one another and not very wide basally, and in the not cylindricnl l)ut much llattened and thorny arm spines. A. vadlcola differs from A. ecnomiotata Clakk in the exceedingly long arms, in the shape of the dorsal and ventral arm plates, and in th<3 f(iW(îr arm spines. Aiujß/tiura ecnoniiotata Clark. Ampkiura ccnomiofatn : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, i). US, fig. 51). Off' Seno-umi, »Suruga Gulf; 31 41 fathoms (Ci.AiiK). Opliiocentrus verticillah^s (1)öderli]in). OpMocnida vertlcUlata : Dodiîrlein, Semon-Zo(j1. Forschuiigsr. Ausb'. 214 ART. 2. — H. MATSÜMOTO u. Malay. Ardi., 1896, p. 287, PI. XIV, figs. 2a & b, PI. XV, figs 7 & 7c. Kœhlee, Exp. Sib(;ga, XLV, Pt. 2, p. 29, PI. II, fig. 4. One specimen ; Enoshima. One specimen ; perhaps Misaki. Malaysian waters. The first specimen is about 5 mm. across the disk, and has eight arm spines in the basal joints. The second is about 12 mm., and has only seven arm spines notwithstanding its larger size. The arm spines decrease in number rather rapidly out- wards, there being only about five in the middle part of the arms. They are not so flattened as in Dödeelein's type, and are rather thorny at the end as in Kœhlee's The disk is covered, between the spines, by a thick skin ; but on drying, the scales around the radial shields be- conie visible, so that the disk is like that of an Opino - nephthys, if we leave the spines out of consideration. The colour in alcohol of the larger specimen is dark olive green on the disk and light yellow Fig. 59. Ophiocerdrus verticUlatus. x 8. a. From above, h. From below, c. Side view of two arm joints near disk. J. Side view of two arm joints in the middle part of an arm. on the arms. Family 3. Ophiotrichidse Ljungman, 1867. Disk covered with plates or scales, or by a naked skin, very MONOGRAPH OF JAPANESE OPHIUROIDEA. 215 often beset witli minute spines or stumpy tubercles. Radial shields very large, with a conspicuous articular socket on the ventral surface near the outer end, fitting to the large, ball-like articular condyle of the genital plate. Genital plates firmly fixed to the basal vertebras. Genital scales short, wide, flat, articulating with the genital plates near the outer end of the latter. Besides, there occur a pair of short, flat scales, just outside each oral shield, supporting the proximal abradial border of the genital slits. Peristomial plates small, entire. Oral frames very stout, with well developed lateral wings for the attachment of voluminous masticatory muscles. In internal view the oral and dental plates are very stout, together forming an X- shape. Teeth very stout, widened, squarish. Dental papillœ well developed, forming a vertical clump at the apex . of each jaw. Oral papilLT absent. Arms inserted ventrally to the disk, horizontally flexible or capable of coiling vertically. Dorsal side of the vertebra) Y-shaped, being con- spicuously notched inwards. Vertebral articulation zygospondyline, the articular peg being always present. Arm spines either moderate- ly long, conical and opaque, or very long, flattened, serrate and hyaline. Tentacle pores large, with one or two scales or none. This family includes fifteen genera, as follows. Ophlothrix MüLLEE Mothridc hirsuta Müller & Troschel. OpMothrix Mrsuta : Müller & Troschel, Sys. Ast., 1842, p. Ill ; yon Martens, Arch. Naturg., XXXVI, 1870, p. 255; Lyman, Rep. Cliallenger, V, 1882, p. 226; Marktanner- Fig. 62. Ophiothrix oMusa. Dorsal view of three arm joints somewhat near disk. x7. 226 ART. 2. — H. MATSUMOTO TuRNERETSCHER, Aim. K. K. Natiu'li. Hofmus. Wien, II, 1887, p. 311, PI. XIII, figs. 34 & 35 ; Kœhler, Bull. Sei. Fr. Belg,, XXXT, 1898, p. 96 ; Ludwig, Abh. Senckenberg. Naturf. Gesell., XXI, 1889, p. 549; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 93 ; Kœhler, Bull. 8ci. Fr. Belg., XLI, 1907, p. 333. Ophiothrix cheneyi : Lyman, Proc. Boston Soc. Nat. Hist., VIII, 1861, p. 84^^; Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, p.. 175. Ophiothrix vo.riabilis : Duncan, Jour. Liun. Soc. Loudon, XXI, 1886, p. 99, PI. XIII, figs. 44 & 35. One specimen ; Pinnacle, Riu-kiu. The specimen at hand appears to be- long to the variabilis-type, having the radial shields almost free of tubercles. Fig. 63. ophiothrix hlrsu- tii: Dor.^al view of an arm base, x 6. Ophiothrix punctoliuibata von Martens. Ophiothrix punctolimhata : VON Martens, Arch. Naturg., XXXVI, 1870, p. 257 ; liYMAN, Ptep. Challenger, V, 1882, p. 227 ; Studer, Abh. K. Preuss. Akad. Wiss. Berliu, 1882, p. 26; Bell, Eep. Zool. Coll. Alert. 1884, p. 143; Bell, Proc. Zool. Soc. London, 1888, p. 388 ; Broc;k, Zeitschr. wiss. Zool., XL VII, 1888, p. 512 : Döderlein, Zool. Jahrb. Sys., ILL, 1888, p. 416, PI. XV, fig. 2 ; LoRiOL, Kev. Suisse Zool, I, 1893, p. 416, PI. XV, fig. 2; Bell, Proc. Zool. Soc. London, 1894, p. 397 ; DÜDERLEIN, Semon— Zool. Forschungsr. V, 1897, p. 294, PI. XIV, figs. 7a & 7b, PI. XVI, figs. 18 k 18a ; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 93. Ophiothrix hirsuta var, iDundolimhata : Marktanner-Turneretscher, Ann. K.K. Naturh. Hofmus., II, 1887, p. 312. Fig. 64. OxMothrix puncto- limhata. Dorsal view oE four arm joints somewhat near disk, x 8. 1) Tliis pajier was not seen by me. MONOGRAPH OF JAPANESE OrHIUROIDEA. 227 One specimen ; locality nnknown, perhaps Misaki. Indo- Pacific. Opli fothrix longipeda (Lamarck). OpMura longipeda : Lamarck, Hist. Nat. Anim. sans Yert., II, 1R16, p. 544. Ophiothrix longipeda : MÜLLER k Troschel, Sys. Ast., 1842, p. 113 ; Lyman, 111. Oat. Mns. Comp. Zool., I, 1865, p. 176; von Martens, Arch. Natiirg., XXXVI, 1870, p. 254; Lyman, Kep. Challenger, V, 1882, p. 220, PL XLVir, %. 4 ; Studer, Abh. K. Prenss. Akad. Wiss. Berlin, 1882, p. 26 ; Marktanner-Turneretscher, Ann. K. K. Naturli. Hofmiis. Wien, II, 1887, PL XIII, fig. 27; Bell, Proc. Zool. Soc. London, 1888, p. 388; Brock, Zeitschr. wiss. Zool., XL Vu, 1888, p. 512 ; Loriol, Kov. Suisse ZooL, I, 1^93, p. 415; Bell, Proc. Zool. Soc. London, 1894, p. 395; DÖDERLEIN, Semon-ZooL Forscliungsr., V, 1899, p. 293, PL XIV, figs. 6a- 6c, PL XVI, figs. 17 & 17a; Kœhler, Bull. Sei. Fr. Belg., XXXI, 1898, p. 97; Ludwig, Abh. Seückenberg. Naturf. Gesell., XXI, 1899, p. 550; Peeffer, ibid., XXV, 1900, p. 85; Kœhler, Exp. Siboga, XIV, Pt. 2, 1905, p. 92 ; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, p. 334 : Clark, Bull. Mus. Comp. ZooL, LI, 1908, p. 298 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 263. ()ne specimen ; Kominato, Bôshù. Two specimens ; Tanabé Bay, Kii. Indo-Pacific. So far as known, Kominato is the northern limit of this species. OpJiioma^a Jaaiekof, sp. nov. One specimen ; Shimabara, Hizen. Fig. 65. Ophiothrix loiwßpeda. Dorsal view of an arm base. xO. 228 AET. 2. — H. MATSUMOTO Diameter of disk 15mm. Length of arms 35 mm. Width of arms at base 1.8 mm. Disk rather flat, covered by the large radial shields and rather coarse, partly imbricated and partly tessellated scales. The scales are covered by skin, so that they are very inconspicuous, especial- ly inside and outside the radial shields, where they are finer than between the radial shields. There are two or three rows of them in each interradius, whore they are coarse, irregular, polygonal, and tessellated. Eadial shields large, triangular, with acute inner angle, longer than two-thirds the disk radius, twice as long as wide, those of a pair separated from each other by a single row of oblong, quadrangular scales, three to five in number. Interbrachial ventral surfaces covered by a soft, naked skin. Genital slits large. Oral shields small, quadrangular, or rather pentagonal, very wide and short, more than twice as wide as long, Adorai shields very small, quadrangular, joined within. The oral and adorai shields have chagreened external surfaces, which are covered, by skin. Buccal Fig. 66. OpMomam kmwkoi. x4. a. From above, h. From below, c. Side view of three .arm joints near disk. with a very large inner angle arONOGRAPH OF JAPANESE OPHIUROIDEA. 229 pores very small. Dental papillœ about twenty-five in number, forming a vertical, oval clump ; in the lower half they are very small and arranged in four rows ; in the upper half they are larger and arranged in two or three rows. Three teeth, quadrangular, with truncated ends ; the lowest one is about as long as the upper dental papillär ; the others are much longer than the lowest, the uppermost one being the longest and narrowest. Arms slender and short, vertically coiled, especially in the distal parts. Dorsal arm plates large, quadrangular, with convex outer border, about twice as wide as long, wider without than within, very convex, so that the dorsal side of the arm is strong- ly keeled as a whole. Four or five basal dorsal arm plates are smaller and continued on to the radial row of disk scales. Some of the dorsal arm plates are divided into equal or unequal halves. Lateral arm plates not very prominent. First ventral arm plate not very small, quadrangular, with curved inner and outer borders and rounded angles, slightly wider than long, wider within than without. Next one or two plates quadrangular, with curved outer and concave lateral borders and rounded outer angles, slightly longer than wide. Those beyond quadrangular, with concave inner and outer Ijorders and rounded outer angles, wider than long. The surface of the ventral arm plates are chagreened and covered by a skin. The successive ventral arm plates are not in contact, but are separated l)y a naked, depressed, groove-like space. Arm spines six or seven in proximal free joints, conical, terete, tapered, blunt ; the uppermost one is short, slightly longer than the corres- ponding arm joint ; the next is the largest, longer than twice the corresponding joint ; the others become shorter downwards, the second lowest being slightly longer than the corresponding joint ; the lowest spine is very small and short, serving as a tentacle 230 ART. 2. — H. MATSUMOTO : scale. The spines rapidly become fewer and shorter outwards, so that in the tenth free arm joint, they are four in nnmber, the uppermost two spines,- which are the longest, being about one and a half times as long as the corresponding joint. In the distal parts of the arms, the lowest spine is transformed into a compound hook. Colour brownish black in alcohol, formerly preserved in formalin. The present species differs from 0. cacaotica Lyman in the more numerous disk scales, in the absence of a distinct central rosette of the primary plates, in the not regularly arranged marginal disk scales, in the narrower arms, in the concave outer border of the ventral arm plates, and in the more numerous arm spines ; from 0. mœrens Kœhler in the larger radial shields, in tlie shape of the oral shields, in the narrower arms, in the wider ventral arm plates, of which the inner and outer borders are concave, and in the more numerous and longer arm spines ; and from 0. obscurci (Ljungman) in the fewer rows of the dorsal inter- radial disk scales, in the larger radial shields, in the narrower arms, in the markedly narrowed basal dorsal arm plates, and in the more numerous arm spines. Ojfhi'otJiela danw Vereill. Ophiothela dance : Verrill, Proc, Boston Soc. Nat. Hist., XII, 1869, p. 391^5; Lyman, Rep. Challenger, V, 1882, p. 230 ; Marktanner-Tueneretschee, Ann. K. K. Naturh. Hofmus. Wien, II, 1887, p. 313 ; Döderlein, Semon — Zool. Forschungsr., V, 1897, "p. 297, PI. XVn, figs. 25-25b ; Kœhler, Ball. 1) This paper was not seen by me. MONOGRAPH OF JAPANESE OrHIUllOlDEA, 231 Sei. Fr. Belg., XXXI, 1898, p. 89 ; Kœhler, Exp. Siboga, XLY, Pt. 2, 1905, p. 117 ; KcEHLER, Bull. Mus. d'Hist. Nai. Paris, 1905, p. 459. Opldothcla isidicola : Lütken, Bull. Hoc. Pioy. Copeuhague, 1872, p. 92, PI. I &, II, figs. 4a.-4g ; Lyman, loc. cit., p. 231 ; Brock, Zeitsclir, wiss. Zool., XLYI, 1888, p. 537; Loriol, Mem. Soc. Phys. d'Hist. Nat. Geuève, XXXn, 1894, p. 52. OpJnothela verrilli : Duncan, Journ. Liuu. Soc. London, XIV, 1878, p. 477, PI. XI, fig. 33. OpMofJiela dance var. mvoluta : Kœhler, loc. cit., 1898, j)- 89. Numerous specimens, clinging to Melitodes ; Moroiso, Misaki ; 5-10 fathoms- Numerous specimens; off Misaki. Strait of Formosa (Lütken). Korea seas (Duncan). Indo-Pacific. There is no doubt, that 0. dance, 0. isidicola and 0. verrilli are conspecific, as they are connected by an unbroken series of intergrading forms between dance and isidicola on the one hand and between isidicola and verrilli on the other. The first of the two lots above men- tioned are mostly of the dance-type, while the second is mostly of the isidicola- but partly of the verrilli- type, both lots con- taining some interme- diate forms. Those of the dance-type are usually dark coloured, being green, blue, purple, or brown, and are annulated on the arms ; those of the isidicola- and verrilli-tjpe are light coloured or white, being light yellow or Fig. 67. Ophiothela dami'. From above. x20. 232 ART. 2. — H. MATSUMOÏO : very light pink. The verrilli-type is closely similar to the isidicola- type, so that the two are practically one. I imagine that on an average, the isidicola-type (including the verrilli-type) has greater bathymetrical range than the dance-type. In some specimens, there are as many as seven arms. Again, a few of the disk grannies may be conical and pointed. In some of the dance-type, the dorsal surface of the disk is entirely free of erranules or tubercles. Order iv. chilophiurida Matsumoto. Disk covered with scales or plates, often with superficial granulations. The radial shield and genital plate on the same side of a radius articulate with each other by means of two articular condyles and one pit on both the plates. Peristomial plates double or triple, not very large. Oral frames witli or without well developed lateral wings. Oral papillae well developed, close-set, very often entirely closing the oral slits ; the outermost one being pointed inwards above the next papilla which is the largest, but sometimes modified in certain ways by the displacement of the second oral tentacle pores outside the oral slits. Arms only horizontally flexible, the vertebral articulation being zygospondyline. Arm plates usually all well developed. Arjii spines very short or long, lying flat on the arm or erect. Key to families of CMlophhtrida, A — Arm spiues short, lying flat on the arm. « — Disk usually free of granules; second oral tentacle pores opening MONOGRAPH OF JAPANESE OPHIÜROIDEA. 233 either outside or inside the oral shts ; arms short and stout, stoutest at the base OpMolepididœ. act — Disk covered with granules ; second oral tentacle pores opening always within the oral shts; arms moderately or very long. b — Arms slender, inserted ventrally to the disk, stoutest at the base or at a distance from it ; arm spines few OpJiioîeucidœ. lb — Arms stout, inserted laterally to the disk, stoutest at the base ; arm spines numerous OpJiiodcrmalidce, pars. AA — Arm spines long and erect. c — Dental papillae absent. d — Disk covered with granules ; oral angles granulated ; arms stout, stoutest at the base ; arm spines numerous .... 0}dModermatidœ, pars. dd — Disk usually free of granules ; oral angles not granulated ; arms slender, stoutest at a distance from the base ; arm spines few .... OphiocMtonidce. cc — -Dental papillae well developed, forming a vertical clump at the apes of each jaw ; arms stout, stoutest at a distance from the base .... Opliiocomidœ. Family 1. Ophiolepididae Ljungman, 1867. Disk covered with thick scales or plates, among which the primaries are often very prominent. Radial shields usually stout. The radial shield and genital plate on the same side of a radius articulate with each other by means of two articular condyles and one articular pit on both the plates. Genital papillae very often, and arm combs sometimes, present. Oral papillœ usually few, arranged in a single series ; where the second oral tentacle pore opens within the oral slit, the outermost oral papilla is pointed inwards and the next one is the largest. Teeth arranged in a single vertical row. Dental papillœ absent. Peristomial plates small, usually double, rarely simple. Arms inserted laterally to 234 ART. 2. — H. MATSUMOTO : ' the disk, short or moderately long, very stont, widest at the base, rapidly tapering outwards, only horizontally flexible. Arm plates either all well developed, or the dorsal and ventral ones may be very rudimentary. Arm spines short, lying flat on the arm. Tentacle scales variable in number, leaf-like, rarely absent. This family is very extensive and includes thirty -eight genera, which may be grouped into two subfamilies as follows. Subfamily 1. Ophiomastinœ Matsumoto, 1915 : — Second oral tentacle pores opening outside the oral slits. I. Several basal lateral arm plates extraordinarily widened ; genital slits invisible or very insignificant ; disk covered merely with primaries and radial shields. Astrophlura Sladen, 1878. Ophiophycis Kœhler, 1901. Ophiomisidium Kœhlee, 1914. Ophiotypa Kœhlee, 1897. OphiomastiK Lyman, 1878. II. Basal lateral arm plates not much widened ; genital slits usually large ; disk covered usually with both primaries and secondary scales, besides the radial shields. Haplophiura Matsumoto, 1915. AntJiopJdura Clark, 1911. Aspklophiura Matsumoto, 1915. Ophlopyrgus Lyman, 1878. Stegophiura Matsumoto, 1915. Amphiophiura Matsumoto, 1915. Gymnophiura Lütken & Moktensen, 1899. Opliiosteh'a Bell, 1902. Ophiochrysis Koehlek, 1904. Homalopldura Claek, 1915. MONOGRAPH OF JAPANESE OPHIUKOIDEA. 235 Ophiiwa Lamaeck, 1816. OpliioperJa Kœhlee, 1912. Ophionohis Bell, 1902.'^ Ophiofjcdfa Moetensen, 1913. Ophiurolepis Matsumoto, 1915. Ophiogona 8tudee, 1878. Ophioplintlms Lyjiax, 1878. Ophlopleura Danielssen, 1877. Ophioden Lyman, 1854. Subfamily 2. Opliiolepidince mihi, 1915 : — Second oral tentacle pores opening within the oral slits. I. Tentacle pores limited to a few basal arm joints. Ophiomusimn Ly:\L\N, 1869. Ophiolipus Lyman, 1878. XL Tentacle pores present throughout the entire length of the arm. Ophiophyllum Ly]vian, 1878. Ophiopenia Claek, 1911. Ophiocra'es Kœhlee, 1904. Ophiomidas Kœhlee, 1904. AmphipholizGna Claek, 1915. Ophiozonella Matsumoto, 1915. Ophiosonoida Claek, 1915. Ophiozona Lyiian, 1865. Ophiothyreus Lytvian, 1865. Ophiolepis Müllee & Teosghel, 1840. Opliioplocus Lyi^lan, 1861. 1) Ophiogly})ha hexadii Smith, 1870, belong-* ^^ ßiy opinion to Ophionolufs, .is distinct from the genuine Ophiur.i. 236 ABT. 2. — H. MATSÜMOTO : Oph iocer amis Lyivian, 1865. The internal structures of the Ophiomastinœ are very divergent, but roughly three types may be distinguished according to the charac- ter of the peristomial plates. The first type, represented by Astro- phiura and Ophioplinthus, lacks the peristomial plates ; the second, represented by Ophiomastus and Opliioden, has simple peristomial plates ; and the third, including the majority of the present sub- family, has double peristomial plates. The internal structures observed in Astrophiura kawamurai are most peculiar. The oral frames and plates are very long and slender, while the dental plates are stout. The peristomial plates are entirely absent, though the peritoneal membranes of the oral region, as well as of the other parts, contain very fine, trans- lucent, perforated scales, much resembling the perforated spicules of holothurians, as seen under a compound microscope. The genital plates and scales are entirely internal, both being very slender and narrow ; the former are very short, with the outer end forming a simple articular face for the radial shield ; the latter are about twice as long as the former, parallel to, and directed above, the adorai shields, not laterally but terminally articulated with the genital plates. The proximal parts of the first lateral arm plate curve inwards and downwards, passing below the radial shield and the genital plate, and articulate with the first vertebra. The dorsal side of each vertebra is not rhomboidal but almost quadrangular, with a rather shallow median groove ; that of the first vertebra is much wider than in the others, the abradial peripheries being very thin and translucent. The ventral side of each vertebra has a rather shallow median groove and a well marked median suture. The vertebra3 of the free arms are almost entirely divided into halves by a very narrow moniliform slit, the MONOGKAPH OF JAPANESE OPHIUROIDEA. 237 halves being very slender and lying very closely side by side. The articulation of the vertebra' is primitive and zygospondyline. The articular peg is situated between the halves of the articular shoulder, the three being almost parallel. In the inner articular face, there is present a well developed articular umbo, the articular knobs being, however, very rudimentary, represented merely by the projecting edges of the central ridge, on which the articular umbo and the pit for the articular peg are placed. Astrophiura appears to remind us more or less of OphiopUnthus meclusœ Lyiian, the internal structures of which have been described and figured by Lyman, by its long oral frames, the absence of distinct peristomial plates and the quadrangular dorsal side of the basal vertebrcT. According to Lyman, OpJuomasfus secundus Lyman has very long oral frames and plates, a small, simple peristomial plate in each interradius and quadrangular, not discoidial, basal vertebrae. I suppose that, the internal structures of Ophiophycis, Ophiomisidium and Ophlotypa are probably similar to those of such genera as Astrophiura, OphiopUnthus and Ophiomastus, though they have not been thoroughly studied. Haplophiura gymnopora (Clark) has also simple, transversely bar-like peristomial plates and short, discoidal basal vertebra?. This genus, as well as Ophiomisidium, Ophiophycis and Astrophiura, lacks the genital bursa? and visible genital slits. Ophiocten sericeum Lyiian is described by the author to have also simple peristomial plates and short, discoidal basal vertebra?. Stegophiura sladeni (Duncan), as well as St. sterea (Claek), has double, transversely bar-like peristomial plates and very high outer ends of the oral frames. The internal structures of Ophiura kinbergi are almost similar to those of Stegophiura except the oral frames, of which the outer ends are not very high. Ophiura 238 ART. 2. — H. MATSUMOTO : /^/w«?^/ (Ljüngman) and Opliiuroleph carinata {HTiwi:n){=: Ophiogli/pha deshayesl Lyman) appear to be almost similar to Ophiura kinbergl in their internal structm-es. Those forms with double peristomial plates have always short oral frames and short, discoidal basal vertebra?. The Ophiolepidince are fairly uniform in their internal' struc- tures, the two opposite extremes being represented by Ophio- mu><'mm and OpJdoceranm. In Ophiomusium trychnum Clark, which is a representative of the species with only two pairs of tentacles to each arm base, the peristomial plates are double and exceeding- ly small, the oral plates are very long and slender, and the oral frames and basal vertebra? are also extremely long. In Ophiomusium cancellatum Lyman, which is a representative of the species with three pairs of tentacles to each arm base, the internal structures are almost similar to those of the foregoing, but the peristomial plates are larger and the oral plates, oral frames and basal vertebra? shorter. In Ophiozonella longispina (Clark), the peristomial plates are comparatively large, the oral plates short and stout, the oral frames more or less short, and the basal vertebra? short and more or less discoidal. In these three species the first vertebra, and not the second, is the shortest. In Ophiomusium, the teeth are very slender, while in Ophiozonella, they are more or less squarish and stout. In Ophiozona impressa . (Lütken), Ophiolepis cincta Müller & Troschel and Ophioplocus japomcus Clark, the peristo- mial plates are comparatively large, the oral plates and frames very short and stout, the basal vertebra? very short and discoidal, the second being the shortest, and the teeth more or less squarish and stout. In Ophioceramis januarii (Lütken), the peristomial plates are also double, but very small and transversely bar-like, the oral plates are very short and stout, the oral frames very MONOGEAPH OF JAPANESE OPHIUROIDEA. 239 stout with well developed lateral wings, the basal vertebriG very short and discoidal, tho second being the shortest, and the teeth squarish and very stout. In all these representatives of the Ophiolepidince, the genital plate and radial shield articulate with each other by means of two condyles and one pit on both the plates, and the genital plate and scale articulate with each other at some distance inwards from the outer end of the latter. Astrophiura Sladen, 1878. Basal portion of arms very intimately united with the disk proper, so as to form with it a pentagonal " asteroid " body, of wliich the dorsal surface is entirely covered over by the stout primaries and radial shields of the disk proper and by the dorsal and much widened lateral arm plates of the much modified arm bases. The interradial borders of the " asteroid " body are hemmed by a series of modified arm spines, which are soldered together. The basal lateral arm plates articulate with the vertebrae by means of a bar-like ridge. Arms -outside the " asteroid " body very abortive, without dorsal and ventral arm plates, without tentacles. Tentacle pores present only within the " asteroid " body, very large and arranged so as to form a pentamerous petaloid series. Ventral arm plates well developed, forming a continuous series, which extends only as far as the last tentacle pores. Oral shield single or five^^; in the former case, it is the madreporite. Adorai shields long and narrow. Oral plates rather stout. A single dental plate is present at the apex of each jaw. »Six or seven oral papilla^ to each jaw, arranged in a single series, 1) A. cavella; Kœhlkk, 1015, has five oral shields. 249 ART. 2. — H. MATSUMOTO : situated deep within the oral shts ; apical two or three of them arise from the dental plate. Teeth and dental papillae proper not present. Interbracliial ventral surfaces covered by a thin skin, which contains fine, hyaline scales. Genital bursa? absent, genital slits invisible. Genital plates and scales entirely internal, very slender, those on the same side of a radius articulating with each other at the end. Peristomial plates absent. Oral frames very long. Vertebra} of the free arms more or less divided into halves by a moniliform longitudinal slit. I do not agree with Sladen in the interpretation of certain structures of the body. Sladen's " side mouth shields " are, in my opinion, genuine oral plates, and his " genital scales " true adorai shields, i. e. side mouth shields. Sladen's " oral papillae " may be retained, if the meaning of these terms be extended to include the scales of the first oral tentacles. Sladen's " apical oral papiUa " or the " single tooth " in my preliminary paper is proved not to be a papilla or tooth, but to be the dental plate. The " septa " or " divisional plates of the ambulacral system " of Sladen are, according to my own observations, not true plates, but merely ridges projecting ventrally from the basal, or adradial, parts of the modified lateral arm plates, and bearing tentacle scales on the maro;in. Astrozyhiuva Jcawamiirai Matsumoto. Astropliiura Iccnvaviîirai : Matsumoto, Auuot. Zool. Japon., Vm, 1913, p. 225, PI. Ill, figs. 1 & 2. One specimen ; Okinosé (a submarine bank), Sagami Sea ; 330 fathoms. Diameter of the pentagonal " asteroid " body 12 mm.; that of MONOGRAPH OF JAPANESE OPHIUEOIDEA. 241 the disk proper, or of the circle passing through the outer ends of the radial shields, 7 ram. Length of the free portion of the arms 6 mm. Width of same near the pentagonal body 0.4 mm. The pentagonal " asteroid " body is very flat, covered entire- ly by the stout primary plates and radial shields of the disk proper and by the dorsal and lateral arm plates of the modified basal portion of the arms. The primary plates are the central, the five infrabasals, five basais, five radiais and five first and five second inter- radials. The infrabasals, basais and interradiais are elevated above the level of the central, radiais and radial shields, so as to form a regular symmetric- al system of ridges. The central plate is ten- sided, the sides corresponding to the infrabasals being longer than those corresponding to the basais. Infrabasals small, rectangular, wider than long. Basais pentagonal, with truncated inner angle, wliich abuts on the central plate, lateral edges concave corresponding to the radiais, much longer than wide. First interradiais rectangular, with concave lateral edges corresponding to the radial shields, much longer than wide. Second interradiais large, pentagonal, with short and Fig. 68. Astrophiura kmcamnrai x6. fibove. b. From below. a. From 242 ART. 2. — H. MATSUMOTO: concave lateral edges corresponding to the radial shields, outer edges very long and concave corresponding to the first lateral arm plates, and terminating with a very acute outer angle. The elevated plates have slightly raised margins, so that the surface is concave. Radiais shaped like maiden hair, with a semicircular inner border and concave outer sides corresponding to the radial shields, and with an acute outer angle ; longer than wide. Radial shields irregularly triangular, with perfectly rounded angles, convex inner and shghtly concave outer borders ; about as long as wide, hardly touching each other. Dorsal arm plates seven or eight in number, confined to the pentagonal body. The first one is triangular, wider than long, smaller than but as wide as the next plate. The following plates, which become smaller outwards, are quadrangular, with truncated outer angles, wider without than witliin ; the second to the fourth plates are wider than long. All the plates are in contact with one another, except the last one or two, which are very rudimentary and isolated. Within the pentagonal body, the lateral arm i^lates are much elongated laterally, with the width almost equal to the length of the corresponding arm joints, lightly bent outwards, striated parallel to the axis of elongation. They meet with the corresponding dorsal arm plates, except the first, which is separated ù'om the first dorsal by the radial shield. The first lateral arm plates of the neighbouring radii are apposed to each other distally, outside the second interradial. The distal edges of the successive lateral plates form a continuous line, the interradial borders of the pentagonal body, which are shghtly indented, and hemmed by a series of soldered papillas, which are evidently the modified arm spines of the modified lateral arm plates. These papillge are longer towards the middle of the interradial borders, where they are MONOGRAPH OF JAPANESE OPHIUROIDEA. 243 about as long as the width of the correspondmg lateral arm plates. There are three or four of them for each lateral plate. With the exception of the peripheral zone, the ventral surface of the interradial spaces of the pentagonal body is covered by a thin skin, which contains fine, close-set, polygonal or rounded, delicate, transparent scales. The peritoneal cavity and the genital glands can be seen through the skin, and apparently extend into the arm bases. The genital bursse are absent and the genital openings are invisible. There is only one oral shield, the mad- reporite, which is irregularly triangular, and so transparent and small as to be almost invisible unless the specimen is dried, and distinguishable only with difficulty from the scales of the inter- brachial ventral surface, and is moreover separated from the adorai shields by a space. The adorai shields are long and very narrow, tapered within to a point, outer end joined to the first ventral arm plate, inner end free, longer than the first ventral plate. Oral plates large and stout, oblong quadrangular, with exceedingly short inner side, convex abradial and concave or nearly straight adradial side, much wider without than within, joined distally to the first ventral arm plate and proximally to each other along the proximal one-third of the abradial side. A single dental plate is present in each jaw, quadrangular in ventral view with curved inner edge. There are six or seven papillae to each jaw, arranged in a horizontal, continuous row, situated deep within the oral slits, and directed not laterally but upwards. The two or three apical ones arise from the dental plate, so that they may be regarded as dental papillae, while the outer papilla3 he just below the first oral tentacle pores and protect the latter, so that they may possibly be the tentacle scales of the same. Ventral arm plates eight or nine in number, limited to within 244 AET. 2. — H. MATSUMOTO : the pentagonal body. The first one is very large, hastate, but with mnch ronnded apex, lateral edges strongly concave corres- ponding to the second oral tentacle pore, and with an indentation at the outer edge, longer than wide, very wide without. Follow- ing plates, which become smaller outwards, quadrangular, with strongly concave lateral edges corresponding to the tentacle pore, with a saddle- shaped surface. They are closely apposed to one another, except the last one or two, which are very rudimentary, rhomboidal and isolated. The tentacles are very large, uniformly diminisliing in size and lying closer together as they proceed outwards, and arranged in a pentamerous petaloid series as a whole. The first pair of tentacles in ventral view are homologous with the second oral tentacles of other ophiurans, because they are situated between the first ventral arm plate and the adorai shields and belong morphologically to the same arm, joint with the latter. They are much larger than the other tentacles. The second pair are the true first arm tentacles. They belong to the same joint with the second ventral and the first dorsal and lateral plates. Each tentacle pore is provided with one or two scales, which are rather small, lanceolate, covered by skin, their concave surface facing the tentacle pore, often turned up. The outermost and smallest tentacle pores have no scale. Each two successive tentacle pores are separated by a ridge of the basal or adradial part of the lateral arm plate, which belongs to the same joint with tlie outer of the two pores. The free distal portion of the arms outside the pentagonal body is very abortive, exceedingly slender, uniformly tapered out- wards, entirely covered by the lateral plates and easy to break. There is a single, exceedingly small and acute arm spine on each lateral plate ; but in the first one or two free joints, there are MONOGKArH OF JAPANESE OPHIUROIDEA. 245 two or three spines. It may be observed in some radii, that the lateral plates of the first free arm joint are in process of being absorbed into the pentagonal body. Colour in alcohol : whitish or light yellow ; central and radiais, radial shields and inner part of the interradial spaces below bluish gray. This specimen was found attached to a stone by the ventral surface of the pentagonal body, in the same way as Chiton or Patella. The present species difiers from the genotype, A. permira Sladen from Madagascar, chiefly in the very regular arrangement of the plates on the dorsal surface of the pentagonal body as a whole, in the larger central plate, in the much smaller and quite regular infi'abasals, in the much narrower basais and interradiais, in the infrabasals, basais and interradiais being elevated so as to form a regular symmetrical system of ridges, in the scarcely joined and not overlapping radial shields, in the absence of the central boss on the primary plates, in the very regular and much narrower dorsal arm plates, in the much smaller and rather incon- spicuous madreporic shield, in the longer adorai shields, in the oral plates, being somewhat narrower and in contact to a less extent, in the perfectly saddle- shaped ventral arm plates and in the much narrower free portion of the arms. Further, this species differs from Chun's Astrophmra from Agulhas Bank, which is not yet named, cliiefly in the smaller and regular infrabasals, in the narrower basais and interradiais, in the elevated infrabasals, basais and interradiais, in having two, instead of three, sets of inter- radials^\ in the not overlapping radial shields, in the absence of 1) The presence of two or three sets of interradials may not be of specific vnhie, because Kœhleb has observed both cases in his ,4. cavellœ. 246 ABT. 2. — H. MATSÜMOTO : the central boss on the primary plates, in the much narrower dorsal arm plates and in the much narrower free portion of the arms. Lately, Kœhlek has described A. cavellœ from the vicinity of the Cape of Good Hope, from which the present species differs chiefly in the mnch shorter and smaller infrabasals, in the much narrower basais and interradiais, in the radiais being not distinct- ly pentagonal but shaped like maiden hair, in the scarcely joined radial shields, in the absence of the central boss on the primary plates, in the much narrower dorsal arm plates, in the much smaller madreporic shield, in the absence of the ordinary oral shields, in the longer adorai shields, in the oral plates being in contact to a less extent, and in the much narrower free portion of the arms. On the whole, A. permira, A. cavellœ and Chun's AstropJiiura (there is some possibility that it belongs to one of the two first mentioned species) appear to me to be nearer to one another than any of them is to the present species. This fact may be correlated with the geographical separation of the present species from the others, which occur near together. Haplophiura Matsumoto, 1915. Disk high, much elevated above the arm bases, covered with rather large plates, among which the primaries are very prominent. Eadial shields stout, joined in pairs. Interbrachial ventral surfaces covered with fine granules. Genital slits indistinct. Oral papillœ soldered together. Teeth in a single vertical row. Dental papillœ absent. Arms short, stout. Arm plates convex. First ventral arm plate larger than tlie following. Few, very small arm spines. Tentacle pores, including the second oral tentacle pore, which opens outside the oral slit, entirely free of scales. MONOGEAPH OF JAPANESE OPHIUBOIDEA. 247 This genus is formed by a single species, Ophiofona gymnopora Clakk, 1909. Revision of Ophiura Lamarck, 18] 6 ( = Ophioglypha Lyman, 1865). As Ophiura s. ext. is a very large and diffuse genus, the necessity for subdivision is admitted by many authors. Ludwig proposed to divide it into two genera according to the presence or absence of the arm combs. But his type of Ophioglyphina is, in my opinion, merely a typical Ophiura. Kœhler tried to divide this genus into two sections according as the second oral tentacle pores opened more or less within the oral slits or entirely outside it. In my opinion, the position of the pores in question is not of such value, though it is of some importance in making subdivisions of Ophiura s. ext. I cannot agree with those authors who try to subdivide this genus by only one or two characters ; all characters should be taken into account. From this point of view, I have come to the conclusion that Lyman's subdivisions may be accepted in principle. I propose here to divide Ophiura, as liitherto understood, into a number of genera, which may be distinguished as follows. A — Adorai shields not oval but long ; no supplementary plates among the oral plates and adorai shields ; genital papillse always, and arm combs usually, present. a — Disk high, covered with very stout plates or very thick scales ; arms higher than, or as high as, wide, also covered Avith very thick, heavy arm plates. • h — Disk covered chiefly or only with very stout primaries and radial shields ; interbrachial ventral surfaces almost entirely occupied by a very stout central plate, besides the very heavy genital scales. 248 ABT. 2. — H. MATSÜMOTO : c— Disk flat ; oral papillse soldered together ; arms consisting of long, knobby joints ; dorsal arm plates rudimentary or absent, ventral arm plates very small, both being widely separated from one another; tentacle pores limited to several basal arm joints, very small; those beyond the second oral pores with a single tentacle scale or without any , Aspidophiura. cc— Disk convex ; oral papillae not soldered together ; arms consisting of short, stout joints; dorsal and ventral arm plates not very rudi- mentary, both being in contact with one anotlier at least in several basal arm joints ; tentacle pores present nearly throughout the entire length of the arms, large, with rather numerous scales AmpMopMura, pars. hh — Disk, as well as interbrachial ventral surfaces, covered with numerous plates or scales ; d — Axms not very short, not very high but cylindrical, gradually tapered outwards, with blunt extremity ; a few, or sometimes numerous, arm spines, arranged in a single row, well spaced .... AmpMopMura, pars. dd — Ai-ms very short, very high, higher than wide, very stout at the base, rapidly tapered outwards, with very acute extremity ; numerous arm spines, very often dimorphic and arranged in two rows • , . . StegopMura. aa — Disk low and flat, covered with rather thin and deHcate plates or scales ; arms low, being much lower than wide, or cylindrical, being nearly as wide as high, covered with not very thick arm plates ; ventral arm plates small, much wider than long, usually separated from one another even in the arm bases ; tentacle pores, except the innermost one or two, very small, with only a few scales HomalopMura & OpMura restr.^^ 1) Claek, 1915, has established Homalophiura, removing a number of species from OpJiiura AS diagnosed in my preliminary paper. The two genera may be distinguished as follows. a' — Disk scales rather few and coarse, the primaries being stout and conspicuous; second oral tentacle pores opening entirely outside the oral sUts ; tentacle pores comaned to two or MONOGRAPH OF JAPANESE OPHroEOIDEA. 249 AA — Adorai shields oval ; one to three supplementary plates present among the oral plates and adorai shields ; genital papiUse and arm combs absent OpUurolepis. Aspidophiura Matsumoto, 1915. Disk rather high, flat, covered chiefly by the very stent primaries and the very large radial shields, which are joined in pairs almost by their entire length. Arm combs present, with conical and acute papillae. Interbrachial ventral surfaces covered with a very large central plate, besides the very stout genital scales. Oral shields large, peculiar in shape, with a beak-Hke inner process. Oral papilla3 soldered together. Arms strongly knotted, with long arm joints. Dorsal arm plates rudimentary or absent. Lateral arm plates well developed, flared ; those of the two sides meeting both above and below. Ventral arm plates small, rhomboidal or triangular. Three arm spines, conical, usual- ly acute. Tentacle pores rather rudimentary, present only in several proximal arm joints. A single tentacle scale to each pore or none beyond the second arm joint. This genus includes Ophioglypha minuta Lyman, 1878, and 0. forhesi Duncan, 1789 {=Ophiura glyptoclisca Claek, 1911), besides the genotype, Aspidophiura loatasei Matsumoto, 1915. This genus stands rather between the solida-groxip of Am- pliiophiura and Anthophiura, but differs from the former in the three basal arm joints ; first ventral arm plate axe-shaped, separated from the next plate Uomalophiura. aa' — Disk scales usually numerous and fine ; second oral tentacle pores opening more or less inside the oral slits; tentacle pores not confined to a few basal arm joints; first ventral arm plate usually wider than long, and usually in contact with the nest plate.... Ophiura, rest. In my opinion. Uomalophiura is close to the Irrorata-gTuow}-) of Ophiura, being however more psedomorphic in certain structures. 250 AET. 2. — H. MATSUMOTO : flat disk, in the peculiar shape of the oral shields, in the soldered oral papillse, in the rudimentary dorsal, as well as ventral, arm plates, in the very small tentacle pores, which are limited to the proximal arm joints, and in the very few tentacle scales ; and from Antliophiura in the entire central plate and in the presence of the genital scales and arm combs. Key to Japanese species of Aspidophiura, A — The six primaries each with a small but conspicuous central boss ; radial shields about as large as the radiais ; ventral arm plates rapidly diminishing in size outwards ; arm spines longer than half the corresponding arm joint ; no tentacle scale beyond the disk . . . . ' loatasei. AA — The six primaries without a central boss ; radial shields distinctly larger than the radiais ; ventral arm plates very slowly diminishing in size outwards ; arm spines shorter than half the corresponding arm joint ; single tentacle scale to each pore beyond the disk .... forhesi. Aspidophiura watasei Matsumoto. Aspidophiura luatasei : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 77. One specimen; off Misaki, Sagami Sea. One specimen; Uraga Channel. Diameter of disk 5 mm. Length of arms unknown, but probably about twice the disk diameter. Width of arms at base 1.3 mm. Disk circular, very flat. The area inside the radial shields is mostly occupied by the six stout primary plates, and the inter- radial spaces of the dorsal surface each by two stout plates. The MONOGRAPH OF JAPANESE OPHIUROIDEA. 251 Fig. 69. Aspidophiura icatasei. x8. a From above, b. From below, c. Side view of three arm joints near disk. pentagonal central, and the hexagonal radiais have each a small bat conspicuous, mamelon- shaped central boss. Central somewhat smaller than the radiais. The first interradiais are pentagonal, or qaadrangular with irregular inner edge, about as long as wide, wider within than without, a little smaller than the radiais, but larger than the second interradiais, which are quadrangular, wider than long, and wider without than within. Radial shields about as large as the radiais, as long as wide, irregularly triangular, with convex abradial side, apposed to each other by their entire length. At the meeting points of the radiais, first interradiais and radial shields, a very small, thick, convex plate is present, and may be duplicated. Radial scales of moderate size, semilunar, with acute, needle-like comb papillae. Interbrachial ventral surfaces entirely covered over by the large, stout genital scales and a very large, stout plate. Genital slits widened at the inner end, bounded by minute, rather well spaced papillae. Oral shields large, stout, triangular as a whole, with widely rounded outer angles and obtuse inner angle, and convex outer and lateral sides, the latter with a blunt notch near the inner end, corresponding to the genital slit ; about as wide as long, wider without than within. Adorai shields narrow, of uniform width, meeting with each other. Five oral papillae on' either side. 252 ART. 2. — H. MATSUMOTO : squarish, wider than long, soldered together. Teeth short, rather triangular, stout. Arms tapered rather rapidly. Dorsal arm plates triangular, with acute inner angle and very convex outer side, about as wide as long, surface very convex. They extend nearly to the extremity of the arms, where they are very rudimentary. Lateral arm plates well developed, conspicuously flared, very convex, meeting both above and below. First ventral arm plate rhomboidal, not very small, as long as wide, inner sides longer than the outer. The second plate, which is the largest, is triangular, with very convex outer side, much wider than long. Those beyond similar to it but smaller, constantly diminishing in size outwards. Three arm spines, needle-like, acute, well spaced ; the middle one longer than the rest, and about two -thirds as long as the corresponding arm joint. Second oral tentacle pore opening entirely outside the oral slit, bounded by about four scales both on the abradial and adradial border. Tentacle pores eight or nine pairs to each arm, very small, destitute of any scales, or a small one may be present on the adorai side in the basal joints. Colour in alcohol : disk yellowish gray above and white below ; arms white. In the specimen from the Uraga Channel, the disk is covered over only by the primaries and radial shields, without any inter- vening smaller scales ; and the central bosses of the six primaries are not very conspicuous. This species differs from A. minuta (Lyman) in the presence of a central boss on the central and radiais, in the well developed arm combs, in the presence of the dorsal arm plates, and in the strongly flared lateral arm plates ; and fi'om A. forbesi (Duncan) in the presence of a central boss on the six primaries, MONOGEAPH Or JAPANESE OPHIUROIDEA. 253 in the relatively larger radiais, in the fewer and smaller intervening scales, in the shape of the first, as well as second, interradials, in the shape of the first ventral arm plate, in the ventral arm plates diminishing in size more rapidly outwards, and in the larger arm spines. Aspidophiura forbesi (Duncan). OpMoglypha forbesi : Duncan, Journ. Linn. Soc. London, XIV, 1879, p. 449, PI. IX, figs. 1-3 ; Lyman, Rep. Challenger, V, 1882, p. 77 ; Kœhlei;, Bull. Sei. Fr. Belg., XXXI, 1898, p. 63 ; Kœhlee, Res. Exp. Siboga, XLV, Pfc. 2, 1905, p. 22. OpJimra glyptodisca : Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 91, fig. 31. • Korean Sea; 51-59 fathoms (Duncan, Clark). Indian Ocean and Malaysian waters. Comparing Duncan's and Clark's descriptions and figures, I can not find any tangible characters, by which Clark's species is distinguished from Duncan's. I therefore look upon the latter as a synonym of the former. The difference in the number of secondary disk scales, which is the only one I can recognise, appears to me to be due merely to the fact, that Clark's type is slightly larger than Duncan's. Stegophiura Matsumoto, 1915. Disk high and arched, covered with thick plates or scales, among which the primaries are conspicuous. Radial sliields stout, more or less joined in pairs. Genital papilla) and arm combs well developed ; papilla) of the latter long, usually spiniform. Oral sliields large, oval or pyriform. Arms very short, stout, higher 254 ART. 2. — H. MATSUMOTO : than wide at the base, tapering outwards very rapidly, with very acute extremity. Dorsal arm plates very well developed, widely in contact with one another. Lateral arm plates high. Ventral arm plates quadrangular, widely in contact with one another. Numerous arm spines, short, spiniform or peg-like, often dimorphic and arranged in two rows. Tentacle pores very large, with very numerous scales. This genus includes Ophiura nodosa Lütken, 1854, 0. stuioitdi LüTKEN, 1857, Ophioglpyha elevata Lyman, 1878, 0. sculpta Duncan, 1879, 0. sladenl Duncan, 1879 {=Ophiura stiphra Claek, 1911) 0. striata Duncan, 1879, 0. sterea Ci-akk, 1908, Ophiura hrachyac- tis Clark, 1911, &c., besides Stegophiura vwipara Matsumoto, 1915, the first species being the genotype. Key to Japanese species of Stegoxjhiura, A. — Arm spines subequal, arranged in a single series. a. — Oral shields very large, almost entirely covering the inter- brachial ventral surfaces, quite reaching to the disk margin ; arm spines minute, flat, fitting into small notches on the inner margin of the succeeding lateral arm plate striata, aa. — Oral shields not very large, not reaching to the disk margin ; arm spines spiniform, acute, free of notches. b. — Radial shields of the same radius overlapping each other ; oral shields pear-shaped ; seven or eight arm spines, including the tentacle scales, which are indistinguishable from the spines vivipara. hh. — Radial shields not overlapping, but joined in pairs ; oral shields pentagonal ; eight arm spines sculpta. A A. — Arm spines dimorphic, arranged in two rows. c. — Arm spines well spaced, not soldered together. d. — Comb papillae not very numerous, short, flat, blunt ; three to six MONOGRAPH OF JAPANESE OPHIUßOIDEA. 255 arm spines, two of wliich are larger and longer than the rest, which are very minute nodosa. dd. — Comb papilla? very numerous, spiniform, acute, but becoming blunter and more jflattened as they pass downwards ; secondary comb papillae i^resent, arising from the dorsal and lateral arm plates of the arm base ; about five primary arm spines, which are more or less erect, and about ten secondary ones, which lie flat on the arm, are present on each lateral plate of the free basal arm joints . - sterea. cc. — Very numerous arm spines, of which the secondary ones, which lie flat on the arm, are soldered together. e. — Three erect primary arm spines, two of which are placed low down and one high up ; between the lower two primary spines lies the lowest secondary spine, which is very large and squarish ; comb papillse very numerous, spiniform, acute, but becoming blunter and more flattened as they pass downwards ; secondary comb papillse present, arising from the dorsal and lateral plates of the arm base ; oral shields ovate sladeni. ee.—Two erect x^riniary arm spines, one of which is placed low down and the other high up ; comb papillse flat and blunt ; oral shields pear-shaped, much longer than wide hr achy actis. Stegophiura striata (Duncan). OpUoglyplia striata : Duncan, Journ. Linn. Soc. London, XIV, 1879, p. 452, PI. IX, figs. 4-5a; Lyman, Rep. Challenger, V, 1882, p. 77. Sondai Bay, Korean Sea (Duncan). Stegophiura vivipara Matsumoto, Stegophiura vivipara : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 79. 256 ART. 2. — H. MA.TSUMOTO ; Nnmerons specimens ; Sagami Sea. Numerous specimens ; Sagami Sea ; 75-100 fathoms. Diameter of disk 6 mm. Length of arms 13 mm. Width of arms at base 1.5 mm. Disk pentagonal or circular (especially when the animal con- tains many embryos), convex, covered with fifty to sixty plates on the dorsal side, including the radial shields. Central plate pentagonal. Five radiais also pentagonal, directly surrounding the central plate, laterally overlapping each other. In each inter- radial space of the dorsal side, there occm^s a large squarish plate, wider than long, and in contact with the radial shields. The latter are irregular in out- line, about as long as wide, one of a pair overlapping the other, instead of being apposed to each other in the radial line. On the ventral side of the disk, the plates are rounded and knob- like, with furrows between ; outer ones larger. Genital papilla? blunt, close-set ; longer outwards and upwards, where they form small arm combs. Oral sliields pear-shaped, much longer than wide, wider without than within, with acute inner angle and perfectly rounded outer border. Adorai shields large, meeting within along their whole width. Five oral papiUae on either side, squarish, very short, wide, close- set. A pair of Pig. 70. Stegophiurci vivipara. x8. a. From above, h. From below, c. Side view of three arm joints near disk. MONCGEAFH OF JAPANESE OPHIUEOIDEA. 257 infradeiital papilloc at the apex of each jaw, much longer and stouter than the other oral papillae and rather obtusely pointed. Five teeth, very small, close- set, obtusely pointed. Arms very short, stout, tapered rapidly outwards. Dorsal arm plates fan-shaped, about as long as wide, convex dorsally. Lateral arm plates convex, those of the two sides separated both above and below in the basal arm joints. First ventral arm plate large, triangular, with obtuse inner angle and convex outer side, wider than long. The following plates are octagonal, with very short inner lateral and outer lateral sides, the former concave at the tentacle pores ; wider than long, wider without than within. From the sixth or seventh outwards, the plates are longer than wide, hexagonal, with very short inner and inner lateral, concave lateral and very convex outer sides. Seven or eight arm spines, including the tentacle scales, in the free basal joints, fine, conical, short ; middle ones longer than the upper and lower ones, and about half as long as the corresponding arm joint ; diminishing in number outwards. The lower spines are much finer and serve as tentacle scales. The second oral tentacle pore is very large, opening outside the oral slit, bounded by three or four scales on both the abradial and adradial sides. The tentacle pores are also large, and bounded in the basal joints by one to three aboral scales, besides the lower arm spines on the adorai side. Colour pale gray in alcohol. The arm length varies from one and a half to two and a half times of the disk diameter. The disk plates, in- cluding the primaries, are often very irregular in size and arrangement. One monstrous specimen has six arms, of which two arise from a common base, corresponding to a pair of radial shields. 258 ART. 2. — H. MATSUMOTO : This species is viviparous. I once dissected out twenty-four embryos of various sizes from a single adult. Stegophiura sculpta (Duncan). Ophioglypha sculpta : Duncan, Journ. Liun. Soc. Londori, XIV, 1879, p. 455, PL IX, figs. 6-8, PI. XI, fig. 35 ; Lyman, Eep. Challenger, V, 1882, p. 77. Oplmira sculpta : Claek, BuU. U. S. Nat. Mus., LXXV, 1911, p. 73. Korean Straits; 23 fathoms (Duncan). Eastern Sea; 95-139 fathoms (Claek). Off Honshu ; 31-41 fathoms (Claek). Stegophiura nodosa (Lutiœn). Opidura nodosa : Lütken, Vid. Meddel., 1854, p. 6; LÜTken, Addit. ad Hist. Oph., I, 1858, p. 48, PI. II, fig. 9 ; Gedsg, Panna Arctica, I, 1900, p. 263 ; Ckirk, BuU. U. S. Nat. Mus., LXXV, 1911, p. 70, fig. 20. Ophioglypha nodosa : Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, !>. 49 ; Lyman, Rep. Challenger, V, 1882, p. 78 ; Kcehlee, Bes. Camp. Sei. Monaco, XXXW, 1909, p. 164, PI. XXVI, figs. 7 & 8. Okhotsk Sea ; 40-109 fathoms (Claek). Berincr Sea. Off Alaska. Arctic Ocean. Stegojyhiura sterea (Clark). Opldoglypha sterea : Claek, Bull. Mus. Comp. Zool., LI, 1908, p. 243. OpJdura sterea : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 75, fig. 22. Numerous specimens ; Okinosé (a submarine bank), Sagami Sea ; 83 fathoms. Numerous specimens ; Uraga Channel ; 35 fathoms. One specimen ; Namerigawa, Etcliii. MONOGRAPH OF JAPANESE OPHIUEOIDEA. 259 Uraga Channel, 70-88 fathoms (Claek). Suruga Gnlf; 124 fathoms (Clark). Off Kii ; 191 fathoms (Claek). Sea of Japan ; 70-114 fathoms (Clakk). Off Korea; 82 fathoms (Clark). Off Honshu ; 57-81 fathoms (Clark). The primary arm spines stand out more or less at right angles to the arm axis, and are conical and longer and stouter than the secondary ones ; they are five in the free basal arm joints. The secondary arm spines are very fine, conical, rather well spaced, nine to eleven in number in the free basal arm joints. Both the primary and secondary arm ^.^ ^^ stegophmra spines diminish in number outwards, the latter •'»■^ere«, side view of an arm base, x 8. more rapidly, so that there are only three primary and no secondary spines in the distal arm joints. StegopJiiiirci sladeni (Duncan). OphioglypJia sladeni : Duncan, Jonrn. Linn. Soc. London, XIV, 1879, p. 458, PI. IX, figs. 9-11; Lyman, Rep. Challenger, V, 1882, p. 77. OpUura stiphra : Clark, Bull. U. S. Nat. Mns., LXXV, 1911, p. 82, fig. 25. Numerous specimens ; Okinosé (a submarine bank), Sagami Sea ; 83 fathoms. Numerous specimens ; Uraga Channel ; 35 fathoms. Off Osé Zaki, Suruga Gulf; 65-125 fathoms (Clark). Off Kagoshima Gulf; 103-152 fathoms (Clark). Korea Strait; 66 fathoms (Clark). Sea of Japan ; 44-47 fathoms (Clark). Off Honshu ; 45-70 fathoms (Clark). The primary arm spines are erect, conical, acute, three in number, one just above the uppermost secondary spine, one just above and the other just below the lowest secondary spine, which 260 ART. 2. — H. MATSUMOTO : is extremely wide and squarish. The secondary arm spines, which He flat on the arm, are squarish, flat, close-set, mostly soldered together, smaller upwards, sixteen to eighteen 0^^^y^ in number in the free basal arm joints ; they I M^^:^^ increase in number with the growth of the wm"! m i animal, so that they are fewer towards the WS D S^ extremity of the arm, as expected from the Fig. 72. Stegopkiura l^W Of localised stagCS. sladeni. Side view of DUNCAN doCS UOt mCUtion tllC sizC Of the an arm base, x 5. type specimen, but judging from the magnifi- cation of his figures, the disk diameter must be about 5 mm. Now, my specimens of corresponding size well agree with Duncan's description and figures. In these young specimens, the lowest secondary arm spine is not very distinctly large, and the secondary arm spines and tentacle scales are apparently similar, forming a continuous series. Again, in Duncan's fig. 11, the primary arm spines are shown to stand apart from the series of the secondary arm spines ; but this is probably due to inaccuracy of observa- tion or drawing. As to the specific identity of the full grown specimens at hand with OpJdura stiphra Clakk, there is no doubt. Thus, my conclusion is that, Ophioglypha sladeni and Ophiura stiphra are merely different stages of one and the same species, the former name having priority. JStegophiura hrachyactis (Clark). Ophiura hrachyactis: Clark, Bull. U.S. Nat. Mus., LXXV, 1911, p. 87, fig. 28. Off" southern Saghalin ; 75-100 fathoms (Claek). MONOGRAPH OF JAPANESE OPHIUROIDEA. 261 Amphiophiura Matsumoto, 1915. Disk high and arched, covered with very thick plates or scales, among which the primaries are very conspicuous. Kadial shields stout, joined in pairs. Genital papillas, as well as papillso of arm combs, well developed, either spiniform or squarish. Oral shields very large, oval, pyriform or trefoil- shaped. Arms moderate- ly long, stout, cylindrical or higher than wide at the base, very gradually tapered outwards, with blunt extremity. Dorsal arm plates usually well developed, widely in contact with one another in the free basal part of the arms. Lateral arm plates high. Ventral arm plates of the basal arm joints quadrangular or axe- shaped, in contact with one another. One to numerous arm spines, short, peg-like. Tentacle pores large, with numerous scales. This genus includes (1) species with very large oral shields, which almost entirely cover the interbrachial ventral surfaces, viz. Ophioglypha hullata Wyville Thomson, 1873 (the genotype), 0. convexa Lyman, 1878, 0. abdita Kœhlee, 1901, 0. insolita Kœhler, 1904, 0. improva Kcehler, 1904 ; (2) species with the interbrachial ventral surfaces covered over by a very large central plate and very stout genital scales, viz. 0. solida Lyman, 1883, 0. stellata Studee, 1883 (1882), 0. sciitata Lyman, 1883, 0. paupera Kœhlee, 1897, 0. sorc?^Y<2 Kœhlee, 1897, 0. liherata Kœhlee, 1904, 0. urhana Kœhlee, 1904, 0. remota Kœhlee, 1904, 0. latro Kœhlee, 1904, Ophiura œdiplax CïjX'rk, 1911, 0. pompophora Qi^m^k, 1911 ; (3) species with the interbrachial ventral surfaces covered with many small plates or scales, and with quadrangular ventral arm plates, viz. Ophioglypha sculptilis LvikLAN, 1878, 0. lacazei Lyman, 1878, 0. lapidar ia Lyman, 1878, 0. undata Lyslan, 1878, 0. ponder osa Lyman, 187 8, 0. prisca Kœhlee, 1904, 0. laudata Kœhlee, 1904, 0. distinda Kœhlee, 262 ART. 2. — H. MATSUMOTO : 1904, Ophiura megapoma Clakk, 1911, 0. penichra Claek, 1911, 0. liadra Claek, 1911 ; (4) species with the interbrachial ventral surfaces covered with many small plates or scales, and with axe- shaped ventral arm plates, viz. Ophioglyplia radiata Lyman, 1878, 0. ornata Lyman, 1878, 0. abcisa Lütken & Moetensen, 1899, 0. obtecta Lütken & Moetensen, 1899 ; besides others. As above indicated, the present genus is divisible rough- ly into four sections, of which the first rather approaches Ophiopyrgus, the second Aspidophiura, the third Stegophiura and Gymnopliiura, and the fourth Opliiura s. str. Key to Japanese species of Ani2)Mophiura, A. — Interbrachial ventral surfaces entirely covered over by the very large oral shield ; central rosette of the central plate, radiais and interradiais very conspicuous ; dorsal, as well as ventral, arm plates widely in contact with one another ; three arm spines, shorb, conical, acute convexa. AA. — Interbrachial ventral surfaces not entirely covered over by the oral shields. a. — Each interbrachial ventral surface covered with a very large central plate and very stout genital scales. h. — Disk covered only by the primaries and radial shields; dorsal, as well as ventral, arm plates longer than wide ; a single very short arm spine œdiplax. hh — Disk covered by the primaries and radial shields and a number of secondary scales ; dorsal, as well as ventral, arm plates wider than long ; three peg-like arm spines pompcypliora. aa. — Interbrachial ventral surfaces covered with many small plates or scales. c. — Comb papillœ S(]uarish, flat, close- set. d, — Oral shields not very narrow ; arm spines peg-like or spiniform, not very numerous, not close-set. MONOGRAPH OF JAPANESE OPHIUROIDEA. 263 e. — Single arm spine ; tentacle scales present on both sides of the pore ; oral angles free of granules peniclira. ee. — Four or five arm spines ; tentacle scales present only on the inner abradial border of the pore ; oral angles often beset with a few granules megapoma. dd. — Oral shields very narrow ; arm spines very numerous, squarish, flat, close-set ; genital papillœ present also on the outer abradial border of the adorai shield ; disk plates convex and lumpy ponderosa. cc. — Comb papillse spiniform, acute, well spaced, though they become shorter, wider, blunter and more close-set downwards. /. — Disk covered with numerous small scales besides the primaries, which are separated from one another; oral shields rounded pen- tagonal, with a very acute inner angle ; five well spaced arm spines sculptilis. ff. — Disk covered with a few secondary plates besides the primaries, which form a central rosette ; oral shields trefoil-shaped ; thirteen or fourteen arm spines lopidaria. A'inpJiiojyhiura convexa (Lyman). OpUoglypha convexa: Lyman, Bull. Mus. Comp. Zool,, V, 1878, p. 84, PI. m, figs. 83 & 84 ; Lyman, Eep. Challenger, V, 1882, p. 58, PI. VI, figs. 13-15; Lyman, Bull. Mus. Comp. Zool., X, 1883, p. 243, PI. IV, figs. 40-45; Kœhlee, Bull. Sei. Fr. Belg., XLI, 1907, p. 293; Kœhler, Res. Camp. Sei. Monaco, XXXIV, 1909, p. 149, PI. XXV, figs. 1 & 2 ; Kœhlee, Bull. U. S. Nat. Mus., LXXXIV, 1914, p. 12. Western North Pacific (type locality) ; 2,050-2,300 fathoms (Lyman). Off western Africa. Caribbean Sea. AtnphiopMura œdiplax (Clark). OpUnra œdip>lax : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 74, fig. 21. 264 ART. 2. H. IVIATSUMOTO ; Sea of Japan; 176-245 fathoms (Clark). Ani2)hio2)7iiura pompophora (Clark). Opldura pompophora : Clark, Bull, U. S. Nat. Mus., LXXV, 1911, p. 89, figs. 29 & 30. Off Suno Saki, Sagami Sea ; 83-158 fathoms (Claek). Eastern Sea; 181 fathoms (Clark). AmphioxjJiiura penichra (Clark). Ophlura penichra : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 84, fig. 26. Off KnsMro, Hokkaido ; 175-464 fathoms (Clark). AmpJiiophiura niegaponia (Clark). Ophiura megapoma : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 79, fig. 23. Off Noto, Sea of Japan; 114-163 fathoms (Clark). Aniphiophiura ponderosa (Lyman). Ophioglypha ponderosa : Lyman, Bull. Mus. Comp. Zool., V, 1878, p. 93, PI. II, figs. 52-54; Lyman, Rep. Challeuger, V, 1882, p. 69, PI. VII, figs. 7-9. Ophiura ponderosa : Clark, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 77. Sagami Sea ; 345 fathoms (Lyman). Off Omai Saki ; 475-505 fathoms (Clark). Okhotsk Sea ; 75-440 fathoms (Clark). Alaska. California. MONOGRAPH OF JAPANESE OPHIUROIDEA. 265 AtnjyJiiopJiiiira sculpt il is (Lyman). Ophioglypha sculptilis : Lyman, Bull. Mus, Comp. Zool., V, 1878, p. 84, PI. IV, figs. 115 & 116; Lyman, Kep. ChaUenger, V, 1882, p. 59, PI. VI, figs. 16-18; IvŒHLEE, Ann. Sei. Nat. Zool., Ser. 8, IV, 1897, p. 301; Kœhlek, Ecli. Indian Mus., Deep-sea Opli., 1899, p. 23 ; Kœhler, Bull. U. S. Nat. Mus., LXXXVI, 1914, 1914, p. 24. Ophiocjh/pha variabilis : Lyman, loc. cit., 1878, p. 85, PI. Ill, figs. 70, 78 & 79; Lyman, loc. cit., 1882, p. 60, PI. VI, figs. 10-12. OpJiiura sculptilis : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 77. Off Boshn; 1,875 fathoms (L™an). Off Yaku-shima ; 1,008 fathoms (Claek). Indo-Pacific and Atlantic. AviphiopJiiura lapidar ia (Lyman). Ophioglypha lapidaria : Lyman, Bull. Mus. Comp. Zool., V, 1878, p. 90, PI. in, figs. 67-69; Lyman, Rep. Challenger, V, 1882, p. ßQ, PI. VII, figs. 16-18. Off Omai Zaki ; 565 fathoms (Lyman). Gymnophiura (Lütken & Mortensen, 1899) mihi, 1915. Disk high, covered by a soft skin, but partially with fine scales. Piadial shields small, narrow, bar-like, widely separated from one another. Genital papillœ, as well as papillae of arm combs, short, flat, squarish, close- set. Oral shields not very large, hour-glass-shaped. Arms rather long, stout, cylindrical. Dorsal, as well as ventral, arm plates well developed, widely in contact with one another. Lateral arm plates high, bearing seven or eight short, peg-like spines. Tentacle pores very large, with numerous scales. 266 ART. 2. — H. MATSUMOTO : This genus contains a single species, G. mollis Lütken à MOETENSEN, 1899. Lütken & Moetensen have estabhshed the present genus to include two species, i.e. G. mollis and G. cœhirescens Lütken & Moetensen, 1899. As far as I can judge, the two species are not so closely related to each other as to justify their inclusion in the same genus, unless Ophkira in a very wide sense be made to serve for it. Judging from the original description and figures, G. cœhirescens appears to be conspecific with, or at least very closely related to, Ophiura flagellata, which is evidently a typical member of Ophiura. The genotype, G. mollis, however, rather reminds us of such species as Amphiophiura ponderosa, A. hadra, A. penichra, &c., by its high disk, cylindrical arms and especially by its squarish and flat genital, as well as comb, papillai. The characters, by which G. mollis is distinguished from the above mentioned species of Amj)hio2)lii2ira, are the naked disk and the narrow and bar-like radial shields, which are widely separated from one an- other. I am inclined to look upon the two peculiarities of G. mollis as the essential characters of the present genus. JloïnalopUiuva clasta Clakk. Ophiura clasta : Claek, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 64, fig. 16. HomahpJiiura clasta : Claek, Mem. Mus. Comp. Zool., XXV, 1915, p. 326.^> Off Omai Zaki ; 507-918 fathoms. Ophiura (Lamaeck, 1816) restr. Disk low and very flat, covered with rather small, imbricating 1) Ihmalophiura Clakk, 1915, is stated by C. to include OpMoglyphn biortata Lyman, 1878 (the genotype), 0. confragosa Lyman, 1878, 0. intorta Lyman, 1878, Q. ahyssorum Lyman, 1873, MONOGRAPH OF JAPANESE OPHIUEOIDEA. 267 scales, among which the primaries are usually very distinct. Radial shields more or less separated from one another. Genital papilhv well developed ; arm combs usually present. Second oral tentacle pores opening more or less inside the oral slits, very large, beset with numerous scales, some of which often form a continuous row with tlie oral papilla\ Arms not very stout but flat or cylindrical, uniformly tapered outwards. Dorsal arm plates usually well developed, often widely in contact with one an- other. Lateral arm plates low, bearing three to seven arm spines, which lie flat on the arm. Ventral arm plates small, separated from one another at least outside the disk. Tentacle pores of the first one or two arm joints large and beset with rather numerous scales ; those beyond are very small and have only one or two scales. This genus, as now restricted, includes (1) species with spini- form genital papillao and comb papilla?, such as Asterias opMura Linné, 1758, Ophiura alhida Foebes, 1839, Opliiolepis robiista Ayres, 1851, Opldura sarsii Lütken, 1854, 0. arctica Lütken, 1854, 0. curnea (Saes) Lütken, 1859, 0. aßnis Lütken, 1859, Ophioglypha lütksnl LiTkiAN, 1860, 0. kinbercji Ljungman, 1866, 0. acervata Lyman, 1869, 0. hrevispina Smith, 1876, 0. inermis Litvian, 1878, 0. papillata Lyman, 1878, 0. flagellata Lyman, 1878, 0. imhecilUs, Lyman, 1878, 0. lepicla LiikiAN, 1878, 0. œqualis Lyman, 1878, 0. Ijungmani Lyman, 1878, 0. meridionalis Litvian, 1879, 0. aurantiaca Yeeeill, 1882, 0. maculata Ludwig, 1886, 0. amphitrltes Bell, 1888, 0. indica Beock, 1888, 0. thouletl Kœhlee, 1895, Ophiozona 0. tessellata Veeeill, 1894, 0 inflata Kœhle::. 1898, 0. divisa Lütken & Moktensen, 1899, 0. nana Lütken & Mobtensen, 1899, 0. scutellata Lütken & Mortensen, 1899, 0. frigida Kœhlee, 1900, 0. gelida Kœhlee, 1900, Ophiozona inermis Bell, 1902, Ophioqlypha brucei Kœhlee, 1907, 0. mimaria Kœhlee, 1907, 0. partita Kœhlee, 19Ö7, 0. scissa Kœhlee. 1970, 0. flexïbUis Kœhlee, 1911, Ophiura clasta Ciaek, 1911, and Ophioglypha rouchi Kœhlee, 1912. 268 ART. 2.— H. MATSÛMOTO : capensis Bell, 1905, Ophiura leptoctenîa Clakk, 1911, 0. mlcra- cantlia Claek, 1911, 0. quadrlspina Oi.kkk, 1911, 0. hatliylia Clark, 1911, Ophioden o'ôplax Claek, 1911 ; (2) species with flat and squarish genital papillae and comb papillae, such as Ophioglypha multisplna Ljungman, 1866, 0. lymani Ljungman, 1870, 0. irromta Lyman, 1878, 0. undulata Lyman, 1878, 0. costata Lyman, 1878, 0. albata Lyman, 1878, 0. jejima Lyman, 1878, 0. loveni LY3^L\N, 1878, 0. fraternel Lyman, 1878, 0. rugosa Lyman, 1878, 0. ambigua Lyman, 1878, 0. tenera Lyman, 1883, 0. falcifera Lyman, 1883, 0. verrucosa Studee, 1883, 0. aspera Kœhlee, 1898, 0. plana LüTKEN & MoETENSEN, 1899, 0. ckmeus Kœhlee, 1904, 0. concreta Kœhlee, 1901, 0. mundata Kœhlee, 1906, Ophiura monostœcha Claek, 1911, 0. atacta Claek, 1911, 0. calyptolepis Claek, 1911, 0. cryptolepis Claek, 1911, 0. paucisquama, nov. and others. OpJiionotus, OpUoperla and Ophiotjalfa are very close to the present genus — especially to the ojjhiura-group. Ophionotus may be defined as ophlura-iovms, with supplementary dorsal arm plates ; Opliioperla as the same with superficial granulations on the disk ; and Ophiotjalfa as the same without genital papilla? or arm combs. Key to Japanese species of Oiihiiira, A. — Genital papillse, as well as comb papillae, spiniform, acute, both or at least the latter being well spaced ; second oral tentacle x^ore opening more or less within the oral slit. B. — Arms flattened, much wider than high at the free arm base ; arm spines rather long and stout opJdura-gvom^ (I). a. — Arm combs well developed, rather large in dorsal view. b. — Three arm spines ; oral shields wider within than without, c. — Disk scales naked ; arm spines not spatulate. MONOGRAPH OF JAPANESE OPHIUROIDEA. 269 d. — Disk scales coarse ; comb pax>illa3 vei'j loDg and slender, eight to ten of them being visible from above Idnhe^-gi. dd. — Disk scales fine ; comb papillae very short, twelve to fourteen of them being visible from above sars'ir CO.— Disk scales more or less obscured by a thick, soft 'skin ; lower arm spines of basal arm joints spatulate ßagellata. hb. — Four arm spines ; oral shields wider without than within ; comb papillae short qiiadrisplna. aa. — Ai'm combs rudimentary or entirely wanting, genital papillœ present. €. — Three exceedingly long arm spines ; dorsal arm plates oval, longer than wide ; disk scales fine ; oral shields Avider without than within , oöplax. ce. — Four arm spines, not exceedingly long ; dorsal arm plates more or less quadrangular, much wider than long ; disk scales coarse ; oral shields wider within than without maculata. BB. — Arms cylindrical, about as wide as high "at the free base; three arm spines, minute, peg-like, short, well spaced imbeciUis-grouT^ (II). /. — Disk scales coarse ; radial shields joiued in pairs ; oral shields trefoil-shaped, about as long as wide ; arm bases within the disk especially wide ; first ventral arm plate very large and rhomboidal ; dorsal arm plates separated from one another imbecillis. ff. — Disk scales fine ; radial shields nearly or entirely separated from one another ; oral shields not trefoil-shaped ; arm bases within the disk not especially wide ; first ventral arm plates small ; dorsal arm plates meeting with one another. g. — Radial shields short, only slightly longer than wide ; oral shields pentagonal, wider than long ; uppermost or lowest arm spine longest, middle one shortest, none of them coming up to the length of the corresponding arm joint micracantha. gg. — Badial shields long, about twice as long as wide ; oral shields pear- shaped, much longer than wide; uppermost arm spine longest, 270 ART, 2. — H. MATSUMOTO : lowest one shortest, the former being as long as, or longer than, the corresponding arm joint leptodenia. AA. — Genital papilla3 and comb papillne squarish, flat, short, blunt, close- set ; second oral tentacle pore opening nearly or entirely outside the oral slit ; arms cylindrical, about as wide as high at the base ; arm spines minute, short irrorata-^xoxv^ (HI)- h. — Disk scales naked ; arm spines well spaced. i. — Three arm spines. j. — The two sets of comb papillte on the sides of an arm base not continuous ; oral shields much wider than long. k. — Oral shields wider without than within ; arm spines minute, sub- equal ; tentacle scales very numerous, twelve to fifteen to the second oral tentacle pore, eight to eleven to the first tentacle pore, five to eight to the second tentacle pore, &c irrorata. Mc. — Oral shields wider within than without ; arm spines not very minute, uppermost one longest and lowest one shortest ; tentacle scales rather few, eight or uine to the second oral tentacle pore, four or five to the first tentacle pore, three or four to the second tentacle pore, and one or two to those beyond paucisqiiama. jj. — The two sets of comb x^apillas on the sides of an arm base form- ing a single unbroken series ; oral shields about as wide as long, wider within than without; arm spines sube(]ual ononostœcha. a. — Six arm spines, minute, peg-like, sube(]ual, evenly spaced ; disk scales thin and hard to distinguish ; oral shields pentagonal, slightly longer than wide albata. Jill. — Disk covered by a thick, soft skin, which obscures the underlying scales ; arm spines closely set. I. — Six or seven arm spines ; disk scales rather coarse and thick, radial shields present ; arm combs rudimentary or entirely absent . . calyptolepis, II. — Seven to nine arm spines; disk scales and radial shields almost aborted in adult specimens ; arm combs well developed cryptolepis. MONOGRAPH OF JAPANESE OPHIUROIDEA. 271 Ojthiuva Mnhevgi (Ljungman). OpMoglypha hlnhergi : Ljungman, Öfv. K. Yet. Akad. Förh., XXVII, 1866, p. 116'>; Lyman, Eep. Challenger, V, 1882, p. 38, PL IV, fig. 70; Kœhlek, Exp. Siboga, XLV, Pt. 2, 1905, p. 22; Kœhler, Bull. Sei. Fr. Belg, XLI, 1907, p. 294. OpMoghipha dnensis : Lyman, 111. Cat. Mus. Comp. Zool., VI, 1871, p. 12, PI. I, figs. 1 & 2 ; Lyman, BuH. Mus. Comp. Zool., V, 1878, p. 99 ; Lyman, loc. cit., 1882 ; Dödeelein, Semon - Zool. Forsclmugsr., V, 1896, p. 281, PI. XV, figs. 3 & 3a; Kœhler, BuU. Sei. Fr. Belg., XXXI, 1898, p. 60, PI. IV, fig. 39 ; Kœhler, Ecli. Indian Mus , Shallow-water Oph , 1900, PI. XV, figs. 6 & 7. OpMoglypha ferruginea : Lyman, loe. cit., 1878, p. 68, PI. Ill, fig. 76. Opldura hinhergl : Clark, Bull. U.S. Nat. Mus., LXXV, 1911, p. 37, fig. 9. Nnmerous specimens ; Tsushima. One specimen ; Tomo, Bingo. Two speci- mens ; mouth of Koajiro Bay, Misaki ; 10 fathoms. Numerous specimens ; Mi- saki. Numerous specimens ; Uraga Channel. Off Yokohama; 8-15 fathoms (Li3iax). Inland Sea; 15 fathoms (Lyman). Fig. 73. Opklura kiréergi. x7. a. From above. h. From below, c. Side view of three arm joints near disk. 1) This paper was not seen by me. 272 ART. 2. — H. MATSUMOTO Eastern Sea ; 53 fathoms (Claek). Off Echigo, Sea of Japan ; 70 fathoms (Claek). Indo-Pacific ; ranging southwards as far as Bass Straits. Ojyhitira sarsii Lütken. Opiiiura sarsii: Lütken, Vid. Meddel, Nov., 1854, p. 7; Lütken, Acldit. ad Hist. Oph., I, 1861, p. 42, PI. I, figs. 3 & 4; Grieg, Fauna Arctica, I, 1900, p. 262; Norman, Ann. Mag. Nat. Hist., Ser. 7, XH, 1903, p. 467; Nichols, Proc. Pt. Irish Acad., XXIV, Sect, B, 1903, p. 254 Clark, Bull. U.S. Nat. Mus., LXXV, 1011, p. 37. OpUoglyjjlia sarsii: Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, p. 41 figs. 2 & 3 ; Lyman, Bull. Mus. Comp. Zool, V, 1878, p. 99 ; Lyman, Kep Challenger, V, 1882, p. 40; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, p 296; Kœhler, Eés. Camp. Sei. Monaco, XXXIV, 1909, p. 115, PI. VH fig. 3 ; Kœhler, Bull. Ü. S. Nat. Mus., LXXXIV, 1914, p. 23, PI. I, figs. 5-6 Four specimens ; off Nameri kawa, Etchii, Okhotsk Sea ; 75-109 fathoms (Clakk). Gulf of Tartary; 318 fathoms. Sea of Japan ; 59-428 fathoms (Claek). Hakodate Bay ; 15.5-47 fathoms (Claek). Off Urakawa, southern coast of Hokkaido ; 175-349 fathoms (Claek). Off Kinkwasan ; 57 fathoms (Claek). Uraga Channel ; 70 fathoms (Claek). Off Korea ; 163-335 fathoms (Claek). Eastern Sea; 181 fathoms (Claek). Fig. 74. Ophmra sarsii. x2. a.. From above, h. From below, c Side view of four arm joints near disk. North Pacific. Arctic Ocean. North Atlantic. MONOGRAPH OF JAPANESE OPHIÜROIDEA. 273 In these Japanese specimens, the primary plates are rather small and have a flat surface, so that the disk is very smooth. The plates and scales of the disk, as well as the arm plates, are Hght gray, with whitish peripheries. In one specimen, the primaries are especially dark, so that the disk appears spotted. Ophiura flagellata (Lyman). OpMoglyplia flagellata : Lyman, Bull. Mus. ,Comp. Zool., V, 1878, p. 69, PI. n, figs. 49-51 ; Lyman, Eep. Challenger, V, 1882, p. 52, PI. IV, figs. 16-18 ; Kœhlee, Aud. Sei. Nat. Zool., Ser. 8, IV, 1896, p. 299 ; Kœhler, Ech. Indian Mus., Deep-sea Oph., 1899, p. 18 ; Kœhler, Exp. Siboga, XIV, Pt. I, 1904, p. 56 ; Kœhler, Mém. Soc. Zool. Fr., XIX, 1906, p. 6 ; Kœhler, BuU. Sei. Fr. Belg., XLI, 1907, p. 294 ; Kœhler, Exp. Sei. Travailleur et Talisman, Vni, 1907, p. 261. Gymnopliiura cœridescens : Lütken & Mortensen, Mem. Mus. Comp. ZooL, XXni, 1899, p. 114, PI. VII, figs. 4-6. OpJiiura flagellata : Clark, Bull. U.S. Nat. Mus., LXXV, 1911, p. 60, fig. 15. Three specimens ; Okinosé, off Misaki ; 100 fathoms. Uraga Channel; 70-302 fathoms (Claek). Sagami Sea; 292- 405 fathoms (CLAPtx). Off Piiknzen, eastern Japan ; 182 fathoms (Clark). Off Omai Zaki ; 475-505 fathoms (Claek). Eastern Sea ; 361 fathoms (Claek). Bering Sea and off Alaska (Claek). Indian Ocean and Malay- sian waters (Kœhlee). South-eastern Atlantic (Kœhlee). The arm spines are much flattened, the lowest one being more or less spatulate. The second oral tentacle pore opens in the oral slit, with four or five abradial and three or four adradial scales, the former showing a tendency to form a continuous series with the oral papillœ. There are three adorai scales to the first 274 ART. 2. — H. MATSUMOTO : six or seven tentacle pores and two to those beyond ; of the aboral scales there are three to the first, two to the second, and one to the thnd. All the tentacle scales are flat, thin and leaf- like. I have carefully compared this species with the description and figures of Gymnopliiura cœndescens Lütken & Mortensen, and have come to the conclusion that the two species agree in almost all characters, e. g. the radial shields, arm combs, interbrachial ventral spaces of the disk, oral shields, adorai shields, oral papillas, dorsal as well as ventral arm plates, arm spines, tentacle scales, &g. The only différence between my specimens and the type of G. cœndescens is that in the former the scales are present all over the disk, while in the latter there is a star- shaped naked central space on the dorsal side. But Lyman says that, the disk of the present species is " covered with a thick skin, under which the thin small scales are scarcely, or not at all, distinguishable." Further, Clark remarks that, his specimens " show a most remarkable difference in the calcification of the disk, for while several of them have the disk covering thick, with the scales hardly distinguishable, as in Lyiman's type, others have it much thinner with the scales evident, while still others have a greater or less part of the disk covered by a naked skin, the calcification being confined to the vicinity of the radial shields and to the interradial margins," and that " the gradation between the two extremes is so complete that there can be no doubt that the amount of calcification is an individual and not a specific or even a local matter." I therefore look upon G. ccerulescens as a synonym of the present species. MONOGRAPH OF JAPANESE OPHIUROIDEA. 275 Ophiuva quadrispina Clark. OpMiira quadrispina: Clark, Bull. U.S. Nat. Mus., LXXV, 1911, p. 55, fig. 13. Okhotsk Sea; 109-188 fathoms (Claek). Yezo Strait; 533 fathoms (Clark). Sea of Japan ; 325-428 fathoms (Clark). Berincr Sea. Alaska. Ophiura oöplax (Clark). Opkiocfen oöplax : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 99, fig. 37. One specimen ; Albatross station (number ?). Off Honshu; 94- 507 fathoms (Clark). Sagami Sea ; 292- 614 fathoms (Clark). Eastern Sea ; 369- 440 fathoms (Clark). In my opinion, this species is evi- dently a typical Ophiura, though Clark has referred it to Opliiocten. In the genuine Opliioc- ten, the tentacle pores Fig. 75. Ophiura oöplax. xl5. a. From above, h. From below, c. Side view of two arm joints near disk, cZ. Dorsal view of two arm joints somewhat near disk. 276 ART. 2. — H. MATSÜMOTO : are of nearly uniform size and the tentacle scales few or none throughout, while in the present species the tentacle scales are very numerous in the very basal pores and rapidly diminish in number outwards. I am not able to find any character distin- guishing this species from the genuine Ophiura. Ophiura niaculata (Ludwig). OpUoglypha maculcda : Ludwig, Zool. Jalirb. Sys., I, 1880, p. 283, PI. VI, figs. 11 & 12 ; Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 49, fig. 11. Two specimens (belonging to Mr. H. Asano, Imperial Bureau of Fishery); off Kitami ; 60 fathoms. Bering Sea. Alaska. Ophiura imbecillis (LymxVn). Ophioglypha imhedllis : Lyman, BuU. Mus.. Comp. Zool., V, 1878, p. 78, PL in, figs. 63 & 64; Lyman, Rep. Challenger, V, 1882, p. 46, PL IV, figs. 11-13 ; Kœhlee, Ann. Sei. Nat. Zool., Ser. 8, IV, 1897, p. 303 ; Kœhler, Ech. Indian Mus., Deep-sea Opli., 1899, p. 21. Opldura imhedllis: Clark, Bull. U.S.Nat. Mus., LXXV, 1911, p. 62. Uraga Channel ; 302 fathoms (Claek). Sagami Sea ; 340- 405 fathoms (Litman, Clark). Off Kii ; 393 fathoms (Claek). Off Hiuga ; 437 fathoms(CLAEK). Eastern Sea; 361 fathoms (Claek). Indian Ocean. Malaysian waters. Ophiura tnicracantha Clark. Opliiura micracnntha : Clark, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 47, fig. 10. Two specimens ; Okinosé, Sagami Sea ; 100 fathoms. MONOGRAPH OF JAPANESE OPHIUROIDEA. 277 Uraga Channel; 197 fathoms (Claek). Off Kagoshima Gulf; 152 fathoms (Clark). Off Clotô Is., Eastern Sea; 139 fathoms (Claek). Ophiuva leptoctenia (Clark). Ophiura leptoctenia : Clark, Bull. U. S. Nat. Mus., LXX, 1911, j)- ^1^ fig. 12. Okhotsk Sea; 73-119 fathoms (Clark). Gulf of Tartaiy ; 318 fathoms (Clark). Sea of Japan ; 92-429 fatlioms (Clark). Southward of Hokkaido ; 175-349 fathoms (Clark). Oft' eastern Japan; 191-507 fathoms (Clark). Off" Omai Zaki ; 624-662 fathoms (Clark). Off" Korea ; 335 fathoms (Clark). Bering Sea. Alaska. British Columbia. Oreijon. Washington. Ojyhiura irrorata (Lyman). OpJiioglypJia irrorata : Ly.man, Bull. Mus. Comp. Zool., V, 1878, p. 73, PI. IV, figs. 106-108 ; Lyman, Hep. Challenger, V, 1882, p. 47, PI. V, figs. 7-9. OpJdoglypha orbiculata : Ly>[AN, Bull. Mus. Comp. ZooL, V, 1878, p. 74, PI. IV, figs. 103-105 ; Lyman, Rep. Challenger, V, 1882, p. 48, PI. VIH, figs. 10-12. OpJiioglypJia grandis : Verrh^l, Proc. U. S. Nat. Mus., XLII, 1894, p. 293. OpJiioglypJia involuta : Kœhler, Ann. Sei. Nat. Zool., Ser. 8, IV, 1897, p. 295, PI. VI, figs. 16-18 ; Kœhler, Ech. Indian Mus., Deep-sea Opb., 1899, •p. 15, PI. Vni, figs. 61-63. OpJiioglypJia tumulosa : Lütken & Mortensen, Mem. Mus. Comp. Zool., XXni, 1899, p. 120, PI. I, figs. 8-13. OpJiiura irrorata : Clark, Bull. Ü. S. Nat, Mus., LXXV, p. 62. Okhotsk Sea; 1,800 fathoms (Clark). Oft" Omai Zaki; 624- 278 ART. 2. — H. MATSÜMOTO : 918 fathoms (Claek). YensM Sea; 943 fathoms (Claek). Off Kii ; 649 fathoms (Claek). Off Hinga ; 720 fathoms (Claek). Cosmopohtan. Ophiura lyaiicisqua^ina, sp. nov. One specimen ; off Misaki, Sagami Sea. Diameter of disk 5.5 mm. Length of arms 20 mm. Width of arms at base 1 mm. Disk pentagonal, with nearly straight interbrachial borders, very flat, thin, covered with numerous coarse, irregular, imbricated scales. The central plate is indistinguishable, but the radiais are distinct, and somewhat larger than the other scales. A large, squarish plate is present in each interradius, just touching the disk margin. Radial shields irre- gularly triangular, with acute inner and much rounded outer angles, a little longer than one- third the disk radius, somewhat longer than wide, not in contact with each other. Arm combs very small, with fine, short, blunt, flat, close- set comb papillae, of which about seven are visible from above. Interbrachial ventral surfaces covered also, with coarse, irregular, imbri- cated scales. Genital slits long, with fine, short, blunt, close-set genital papillœ on the abradial border. Fig. 76. Opldura paucisquama. x 10. «. From above. 7;. From below, c. Side view of three arm joints near disk. MONOGEAPH OF JAPANESE OrHIUEOIDEA. 279 Oral shields pentagonal, or rather lyre-shaped, with straight inner and notched lateral sides and a slightly convex outer side, wider than long, widest at the inner lateral angles ; outer angles perfectly rounded. Adorai shields very long, narrow, meeting with each other within, constricted at about the middle by the second oral tentacle pores and the inner lateral angles of the oral shields ; the outer lobe separates the oral shields from the first ventral and lateral arm plates ; the inner lobe is wider within than without. Oral plates parallel to the adorai shields, also long and narrow. Four or five oral papillœ on either side, close-set, inner ones longer, narrower and acuter ; the outermost one is very wide and short. Besides, there occur two or three long, conical, acute apical papillœ in each jaw. Teeth conical, acute, stout. Arms slender, uniformly tapered outwards. The first dorsal arm plate is exceedingly small, situated between the pairs of radial shields and of arm combs ; the second is also rather small. Tlie third is the largest, pentagonal, with very short inner side and rounded outer angles, about as long as wide. The fourth is similarly pentagonal, but longer than wide, and the inner side much shorter. These plates are in contact with one another, but the rest are separated by the lateral arm plates, rhomboidal, with acute inner and rounded outer angles, longer than wide. They constantly diminish in size outwards, and become exceedingly small near the extremity of the arms. Lateral arm plates flared, meeting both above and below, except in the first joint, in which they do not meet below, and the first three or four free joints, in which they do not meet above. First ventral arm plate not very small, hexagonal, with very short lateral sides and rounded angles, much wider than long, about as wide as, but shorter than, the next plate. The latter is the largest, hexagonal, with very short 280 AET, 2. — H. MATSÜMOTO : inner and very convex outer sides ; inner lateral and lateral sides concave, the former being in contact with the lateral arm plates and the latter with the tentacle pores ; longer than wide, wider without than within, in contact with the first plate. The third plate is pentagonal, with concave inner and lateral sides and strong- ly convex outer side, about as wide as long. Those beyond are triangular, with wide inner angle and very convex outer side, wider than long ; separated from one another and constantly diminishing in size outwards. Three arm spines, conical, acute, the uppermost is longer and stouter than the others, and are two- thirds as long as the corresponding arm joints ; in the free basal joints the lowermost two are half as long as the corresponding joints. The second oral tentacle pore opens hardly within the oral slit, with four or five scales on both the abradial and adradial borders. The first four or five tentacle pores have each two scales on the adorai border, but the rest have only one. Besides, the first and second pores have two very small scales on the aboral border, which are almost covered over by the adorai scales. Colour in alcohol : disk very light brown, arms white. The present species differs from 0. irrorata (Lyman) in the shape of the oral shields, in the longer arm spines, and in the fewer tentacle scales ; from 0. plana (Lütken & Moetensen) in the shape of the radial and oral shields, in the coarser scales of the interbrachial ventral surfaces, and in the fewer tentacle scales ; and from 0. m'lndata in the shape of the radial and oral shields, in the longer arm spines, and in the fewer tentacle scales. MONOGRAPH OF JAPANESE OPHIUKOIDEA. 281 Ophiura inonostœcJia Clark. Opliiura monosfœcha : Bnll. Ü. S. Nat. Mus., LXXV, 1911, p. 65, fig. 17. Off Honshu; 153 fathoms (ClaPvk). Suruga Gnlf ; 270 fathoms (Claek). Off Hiuga ; 405-578 fathoms (Claek). Ophiura alhnta (Lyman). Opldoglyplia albcda : Lyman, Bull. Mus. Comp. Zool., Y, 1873, p. 77, PI. IV, figs. 95-97 ; Lyman, Rep. Clialleuger, V, 1882, p. 51, PI. V, figs. 13-15. Sagami Sea ; 775 fathoms (Lyman). Ophiura calyptolepis (Clark). Ophiura calyptolepis : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 67, fig. 18. Uraga Channel ; 70-197 fathoms (Claek). Sagami Sea ; 153- 405 fathoms (Claek). Ophiura cryptolepis (Clark). Ophiura cryptolepis : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 1911, fig. 19. Off Omai Zaki ; 475-505 fathoms (Claek). Bering Sea. Alaska. Washington. Ophiurolepis Matsumoto, 1915. Disk covered with two kinds of scales, the larger ones being surrounded by a zone of smaller ones. Radial shields moderate- ly large, rounded, widely separated from one another. Genital 282 ART. 2. — H. MATSUMOTO : papillœ and arm combs absent. Adorai shields oval, with pointed outer end. One to three supplementary plates present among the oral plates and adorai shields. Oral papillae close-set, completely closing the oral slits. Teeth in a single vertical row. Dental papillae absent. Arms long, stout, with very short arm joints. Dorsal arm plates well developed, wide, strongly convex. Ventral arm plates triangular, nearly in contact with, or separated from, one another. Second oral tentacle pore opening entirely outside the oral slit, long, entirely closed by the tentacle scales, which are so modified as to appear like supplementary plates. A single arm spine and three tentacle scales, both being very small and peg- like. This genus includes a single species, Ophiolepis carinata Stüdee, 1876 (referred to Opliiocjlypha by Studee, 1883, after the publica- tion of Ophioglypha deshayesi by Lyinian, 1878). In my opinion Ophiolepis carinata Studee and Ophioglypha deshayesi Lyman are conspecific, the former having priority. Through Dr. HuBEET L. Claek's kindness, I was enabled to examine one of Lyman's specimens and I am convinced that, it agrees fairly well with Studee's description and figures, though the latter are rather imperfect and inaccurate. Lyman did not take into account Opliiolepis carinata when describing OphioglypJia deshayesi, probably because he took Studee's statement as to genus on trust. It may be remarked that, both Studee's and Lyman's specimens were taken from the vicinities of Kerguelen Island, the former at the depth of 60-65 fathoms, the latter at 28-150 fathoms. Ophiurolepis is not very close to any other subdivision of Ophiura s. ext., though it more or less resembles the irrorata-gaow^ of Ophiura s. str. Ophiurolepis resembles Opliiolepis and Ophiozona restr., merely in the squamation of the disk. MOXOGKAPH OF JAPANESE OPHIUROIDEA. 283 Key to Japanese genera of OpMolejndinœ, A — Tentacle pores limited to a few basal arm joints ; disk plates or scales, as well as radial shields, more or less intimately soldered together; oral papillae more or less soldered together to form a single x>iece ; arm joints long and hom'-glass-shaped ; dorsal and ventral arm plates more or less rudimentary ; arm spines very minute and peg-like Ophiomnsium. AA — Tentacle pores well developed throughout the entire length of the arm ; disk plates or scales imbricating ; oral papillae not soldered together ; arm joints not very long ; dorsal and ventral arm plates usually well developed ; arm spines not very minute. a — No true tentacle scales, though the lower arm spines may some- times serve as one ; disk covered with fine, granule-like scales ; genital shts very minute, placed at the inner corners of the inter- brachial ventral surfaces ; dorsal arm plates more or less widely separated from one another by supplementary plates or naked spaces ; ventral arm plates separated from one another at least in the basal arm joints Ophiopenia. aa — Tentacle scales always well developed. b — Dorsal arm plates entire, without supplementary plates ; ventral arm plates rhomboidal and separated from one another except at the arm base ; lateral arm plates of either side meeting both above and below except at the arm base ; disk covered with large plates and smaller scales ; radial shields very stout ; one or two tentacle scales Opliiozonella. bh — Dorsal arm plates not entire, accompanied by supplementary plates ; ventral arm plates quadrangular, widely in contact with one another ; lateral arm plates of either side not meeting above or below. c — Dorsal arm plates proper not divided, accompanied by small sup- plementary plates on both sides or also along the outer border ; disk covered with two kinds of scales, the larger ones surrounded by a zone of smaller ones ; radial shields moderately stout ; two tentacle scales, together forming an oval OpJdolepis. 284 AET. 2. — H. MATSUMOTO : ce — Dorsal arm plates proper divided into halves, which are widely separated from each other by a mosaic of supplementary plates ; disk covered with very fine scales ; radial shields very small ; three or four tentacle scales Ophioplocus. Key to Japanese species of OiJhioniusiuni, A — Two pairs of tentacle pores to each radius ; genital slits free of papillse ; genital scales extremely stout ; a very conspicuous, large, central plate is present in each interbrachial ventral surface ; disk covered with very stout plates and very large radial shields. a — Genital slits very short and j)oi'e-like, or invisible, free from the oral and adorai shields, which are completely joined ; no ventral arm plates beyond the third. b — Disk margin free of tubercles ; disk with some secondary plates, besides the primaries and radial shields ; two or three arm spines. c — Disk plates more or less imbricated, or arranged like steps scalaire. ce — Disk plates tessellated simplex. hh — Disk margin beset with tubercles ; disk with only tlie primaries, besides the radial shields ; five arm spines granosum. aa — Genital slits long, peculiarly situated between tlie oral and adorai shields, the former being almost axe-shaped. d — No ventral arm plates beyond the third. e — Disk covered with rather small primaries and smaller secondary plates, besides the radial shields, which are separated from one another ; dorsal side of the disk and arms usually very tubercular ; interbrachial ventral surfaces covered with a large central and several secondary plates, besides the stout genital scales. . . .lymani. ee — Disk with only very stout primaries, besides the radial shields, which are joined ia pairs ; dorsal side of the disk and arms smooth ; inter brachial ventral surfaces covered only by a very large central plate, besides the very stout genital scales lunare. MONOGRAPH OF JAPANESE OPHIUEOIDEA. 285 dd — Ventral arm plates present tliroughout the entire length of the arm. ,/ — Disk covered with stout primaries and smaller secondary plates, besides the radial shields ; each primary plate bearing a very large central umbo ; five or six peg-lik arm spines tryclinum. ff — Disk covered only by very stout primaries and radial shields ; no central umbo on them ; three very large, acute, erect arm spines, of which the uppermost and largest is x>laced on the dorsal surface of the lateral arm plate lütkeni. AA — Three pairs of tentacle pores to each radius ; genital slits provided Avith papillae ; genital scales not very stout ; interbrachial ventral surfaces coA'ered with many small plates or scales, without very prominent central plate ; disk covered with many small plates or scales, besides the radial shields. g — Genital slits not reaching to the disk margin ; each tentacle pore provided with one scale on the abradial bjrder and often also with an additional one on the adradial border ; a group of j)apill8e present on either side of each arm base, just outside the genital slit and radial shield ; three close-set arm spines. h — Oral shields triangular, about as wide as long, wider without than within cancellatuni. Idi — Oral shields pear-shaped, much longer than wide, wider within than without .joUiensc. r;^— Genital slits reaching to the disk margin, each tentacle pore provided with two scales on the abradial border ; no papillae on the sides of the arm base ; oral shields triangular, about as wide as long ; four well spaced arm spines laqueatiim. Ojyhiomusiufn scalare Lyäian. OpMomvsium scalare : Lyman, Bull. Mus. Comp. Zool., V, 1878, p. 117, PI. I, figs. 1-3 ; Lyman, Kep. Challenger, V, 1882, p. 95, PI. I, figs. 4-6 ; Kœhlee, Ann. Sei. Nat. Zool., Sér. IV, 1896, p. 308, PI. VI, figs. 24 & 286 ART. 2. — H. MATSÜMOTO : 25 ; Kœhlee, Ech. Indian Mus., Deep-sea Oph., 1899, p. 26, PI. II, figs. 12 & 13, PI. in, fig. 21 ; Kœhler, Exp. Siboga, XLV, Pt. 1, 1904, p. 65 ; Kœhlek, Mém. Soc. Zool. Fr., XIX, 1906, p. 6; Kœhler, Exp. Sei. Travaüleur et Talisman, VIII, 1907, p. 266. Two specimens ; off Ukishima, Uraga Channel ; 300 fathoms. Indo-Pacific. Atlantic. Diameter of disk 9 mm. Length of arms 28 mm. Width of arms at base 2 mm. Disk pentagonal, rather flat. Dorsal surface covered with about sixty-six stout, somewhat imbricated plates, besides the radial shields ; the central plate, basais, first radiais, second radiais, first interradiais and second interradiais being the largest. Central plate pentagonal, with the rounded angles placed interradially. First radiais semicircular or diamond- shaped, wider than long, separated from the central plate by the small, oblong infrabasals, and from one another by the diamond -shaped basais. First interradiais longer than wide, much wider Fig. 77. Ophiomusmm scalare xlO. n. From above, h. From below, c. Side view of three arm joints near at the OUtcr end than disk. at the mner. becon'd interradiais quadrangular, wider than long. Piadial shields large MONOGKAPH OF JAPANESE OPHIUROIDEA. 287 and stout, irregularly triangular, longer than wide, longer without than within ; those of a pair separated from each other by the large second radial, small third radial and the small first dorsal arm plate. Interbrachial ventral surfaces covered chiefly by the very large, stout genital scales and a very large, stout, convex plate, lying directly outside the oral shield. Genital slits small, very short and narrow. Oral shields large, pentagonal, with very long inner and very short lateral sides, an acute inner and rounded outer angles, stout, longer than wide, wider without than within. Adorai shields very large, stout, long, wide, tapered inwards, meeting with each other. Seven or eight oral papillas on either side, squarish, flat, soldered together, and obscuring the original number ; the outermost one is exceedingly large and wide, and arises from the adorai shield ; the next is much smaller and narrower than the preceding, but distinctly larger and wider than the following. Three teeth, small, triangular, pointed. Arms cylindrical, gradually tapered. First dorsal arm plate very small, rounded triangular, convex, wedged in between the radial shields. The next is also small, quadrangular, with rounded angles and convex sides, wider than long, with convex surface, in contact with the first. The five or six remaining are small and triangular, with rounded angles and convex sides and surface, and diminish in size outwards. Lateral arm plates well developed, covering almost the entire surface of the arms, convex, flared. Only three ventral arm plates : the first small, pentagonal, wedged in between the adorai shields ; the second and third also small, but a little larger than the first, pentagonal, with the tentacle pores at the lateral angles. Only two pairs of tentacle pores and scales ; the latter small and oval, the first scales being larger than 288 ART. 2. — H. MATSUMOTO: the second. Two or three arm spines, very minute, peg-like, situated close together in a notch. The plates of the disk and arms are minutely tuberculated. Colour in alcohol white or pale gray. In the smaller one of these specimens, the tentacle pores and scales are entirely absent, so that the second and third ventral arm plates are triangular instead of being pentagonal. Lyman's type was much smaller and Kœhlee's specimens much larger than mine. The disk plates are therefore fewest in Lyman's and most numerous in Kœhlee's ; they are moreover separated from one another by narrow grooves in the type, but imbricated in both Kœhlee's specimens and mine. The radial shields are closely in contact in the type, but separated in my specimens, and more so in Kœhlee's. The oral papillœ are most numerous in mine. The two small plates shown in my figure between the oral shield and the largest interbrachial plate do not occur in the type, and are also absent in Kœhlee's specimens and the smaller one of mine. Both in the type and in my specimens, the dorsal surface of the disk and arms is not so prominently tuberculated as is indicated in Kœhlee's fio-ures. OjyJiiomusiwin simplex Lyiman. OpJtiomusmm simplex : Lyman, BuU. Mus. Comp. Zool., V, 1878, p. 115, PI. I, figs. 10 & 11 ; Lyman, Eep. Challenger, V, 1882, p. 93, PL I, figs. 7-9 ; Clark, Bull. U, S. Nat. Mus., LXXV, 1911, p. 109. OpMomnsinm sanctum: Ivœhler, Exp. Siboga, XLY, Pt. I, 1904, p. 59, PL XT, figs. 7-9. Eastern Sea; 71-130 fathoms (Claek). Amboyna, Malaysian waters. MONOGRAPH OF JAPANESE OPHIUROIDEA. 289 Ophioniusiuin granosuin Lyman. Op7iiomi(sh(m granosum : Lyman, Bull. Mus. Comp. Zool., V, 1878, p. 118, PI. I, figs. 12 & 13; Lyman, Kep. Challenger, V, 1882, p. 96, PI. I, figs. 10-12. Off Boshii; 1,875 fatlioms (Lymak). Ophiomusittin lyniani Wyville Thomson. Ophiomusium lymani : Wyville TnoisrsoN, Depth of the Sea, 1873, p. 172, figs. 32 & 33 ; Lyman, Bull. Mus. Comp. Zool., V, 1878, p. 113 ; Ly^afax, Kep. Challenger, V, 1882, p. 90; Lyman, Bull. Mus. Comp. Zool., X, 1883, p. 245, PI. V, figs. 55-57 ; Bell, Cat. Brit. Ech., 1892, p. 115; Ivœhlee, Eésult. Camp. Sei. Caudan, 1896, p. 72; Kœhler, Mém. Soc. Zool. Fr., IX, 1896, p. 204 & 242 ; Kœhler, Résuit. Camp. Sei. Monaco, Xn, 1898, p. 42; Ludwig, Sitzungsber. Akad. Wiss. Berlin, 1899, p. 220; LÜTKEN & Mortensen, Mem. Mus. Comp. Zool., XXIII, 1899, p. 137, PI. m, figs. 8-11; Kœhler, Exp. Siboga, XLV, Pt. I, 1904, p. 58; Kœhler, Result. Camp. Sei. Monaco, XXXIV, 1909, p. 161, PI. III, fig. 1, PI. TV, fig. 1 ; McClendon, Univ. California Public, Zool., VI, Pt. 3, 1909, p. 36, PI. I, figs. 4 & 5 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 107 ; Mortensen, Meddel. om Grönland, Kj(|)benhaven, XXIII, 1913, p. 354; Kœhler, Bull. U.S. Nat. Mus., LXXXIV, 1914, p. 26. Eastern Sea; 152 fathoms (Clark). Off Ornai Zaki ; 624-914 fathoms (Claek). Off eastern Japan ; 507-720 fathoms (Clakk). Widely distributed both in the Indo-Pacific and Atlantic. Ojyhio^misiuni lunare Ly^fan. OjjMomnsium, lunare : Ly.man, Bull. Mus. Comp, Zool., V, 1878, p. 116, PL I, figs. 4-6 ; Lyman, Kep. Challenger, V, 1882, p. 84, PI. I, figs. 13- 15 ; Kœhler, Esp. Siboga, XLV, 1904, p. 58 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 107. 290 AET. 2. — H. MATSUMOTO : Eastern Sea ; 103-152 fathoms (Claek). Malaysian waters. Ophioniiisiuiii trychnum Claek. Ophiomusmm trycJmum : Clark, Bull. U. S. Kat. Mns., LXXV, 1911, p. 109, fig. 40. Two specimens ; Iwatogaké, Sagami Sea ; 300 fathoms. Two specimens ; Okinosé, Sagami Sea. Three specimens ; locality- unknown. Surnga Gulf; 94-150 fathoms (Clark). Uraga Channel; 70- 302 fathoms (Claek). Off eastern .Japan; 191-578 fathoms (Claek). In my specimens, there are always two, instead of three, very large plates in each dorsal interbrachial space. The radial shields and the large interradial marginal plates are often very rough and tuber- cular. The dorsal side of the arms also shows Fig. 78. Ophiomusmm irychnum. From above, xi. a tendency to be rough and uneven. Ojjhiomusiiitn lïïtkeMl Lyman. Opldomushim liUkeni : Lyman, BuU. Mus. Comp. Zool., V, 1878, p. 114, PI. V, figs. 138-140; Lyman, Eep. ClialleDger, V, 1882, p. 91, PL I, figs. 16-18 ; Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 107. Eastern Sea ; 139-152 fathoms (Claek). MONOGEAPH OF JAPANESE OPHIÜEOIDEA. 291 Arafiira Sea ; 129 fathoms (Li-^iax). Ophioniiisiiim cancellatutn Ly.man. OpUomitsium cancellahün : Lyman, Bull. Mus. Comp. Zool,, V, 1878, p. Ill, PL I, figs. 17 & 18 ; Lyman, Kep. Challenger, V, 1882, p. 88, PL n, figs. 16-18 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 106. Numerous specimens ; Uraga Channel. Colnett Strait ; 1,008 fathoms (Claek). Off Heda ; 168 fathoms (Clark). Suruga Gulf; 65-297 fathoms (Clark). Oif Omai Zaki ; 475-505 fathoms (Clark). Sagami Sea; 369-775 fathoms (Clark, Lyman). Uraga Channel ; 197 fathoms (Clark). Off eastern Japan; 191-578 fathoms (Clark). Off Bermudas (Lyman). OjyUiomnsiiiin jolllense McClendon. OpMomusium JoUietisis : McClendon, Univ. California Public, ZooL, ^T:, Pt. 3, 1909, p. 36, PL I, figs. 2 & 3 ; Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. Ill, fig. 41. Off Heda; 167 fathoms (Clark). Uraga Channel ; 197 fathoms (Clark). Off eastern Japan; 191-505 fathoms (Clark). California. Washing;ton. OpJiioniusium laqiteatiun (Ly^^ian). OpMomusium laqueatum : Lyman, Bull. Mus. Comp. ZooL, V, 1878, p., 113, PL I, figs. 14-16; Lyman, Ptep. Challenger, V, 1882, p. 90, PLU figs. 10-12 ; Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 106. Eastern -Sea ; 103-152 fathoms (Clark). Lat. 5- 42' S., long. 132= 25' E.; 129 fathoms (Ly^l\n). 292 ART. 2. — H, MATSUMOTO : Ophiopenia discicantha Clark. OpJdopenia disacantha : Clark, Bull. U.S. Nat. Mus,, LXXY, 1911, p. 102, fig. 38. Sea of Japan ; 176-200 fathoms (Claek). Bering Sea. Alaska. Revision of Ophiozona. I have great doubt about the generic value of Oplnozona as now understood. The genotype, Ophiolepis impi'essa Lütken, 1859, has very numerous small disk plates surrounded by a zone of much finer scales, and well developed, quadrangular dorsal, as well as ventral, arm plates, which are perfectly joined to one an- other. A characteristic feature is the presence of a group of three disk plates between the outer parts of each pair of radial shields. In every character, the genotype is very nearly allied to the genuine Ophiolepis and distinguished from it merely by the absence of supplementary dorsal arm plates. But most species of Ophmicma as understood at present, differ much from the genotype, having less numerous large disk plates and less well developed, rhomboidal dorsal, as well as ventral, arm plates, which are separated from one another outside the arm base by the lateral arm plates. Moreover, the trio of disk plates above mentioned is entirely absent. Another atypical group of Ophiozona is represented by 0. gymnopora Claek, 1909, which is characterised by the disk being elevated high above the arm bases, by the completely joined radial shields, by the granulated interbrachial ventral surfaces, by the indistinct genital slits, by the first ventral arm plate being larger than the following, by the second oral tentacle pore opening entirely outside the oral slit, and by the total absence of the MONOGRAPH OF JAPANESE OPHIUKOIDEA. 29:> tentacle scales. I propose to distinguish the first atypical group as OpliiozoncUa and the second as Haplophiura from the genuine Ophiozona, the distinctive characters of the three groups being as follows. A — Disk low and flat, not nmch higher than the arm bases ; radial shields separated from one another ; interbracliial ventral surfaces free of granules ; genital slits long ; first ventral arm j)late smaller than the following ; second oral tentacle pore opening within the oral sHt ; tentacle scales present. a — Disk covered with very numerous small plates, of which the larger ones are surrounded by a zone of smaller ones ; a trio of disk plates distinct between the outer parts of each pair of radial shields; arms very gradually tapering outwards, with rather blunt extremity ; dorsal, as well as ventral, arm plates almost quadrangular, widely in contact with one another through almost the entire length of the arm OpJdozona, restr. aa — Disk covered with less numerous large plates mingled with smaller ones ; no trio of disk plates between each pair of radial shields ; arms rather rapidly tapering outwards, with very acute extremity ; dorsal, as well as ventral, arm plates less well developed, separated from one another by the lateral arm plates except at the arm base, where they may be in contact Ophlozonella. AA — Disk elevated high above the arm bases ; radial shields joined in . pairs ; interbrachial ventral surfaces closely covered with granules ; genital slits indistinct ; first ventral arm plate larger than the following ; second oral tentacle pore opening entirely outside the oral slit ; tentacle scales absent. HaplopJmira. There are two species referred to Ophiozona, viz. 0. inermis Bell, 1902, and 0. capensis, Bell, 1905, which are left out of account in the above key. The latter is in my opinion referable to Ophiura s. str., while the former is referred to Homalophiura by Clakk. 294 ART. 2. — H. MATSUMOÏO : Ophiozonella Matsumoto, 1915. Disk covered with stout plates mingled with smaller ones. Radial shields large, usually separated from one another, but sometimes slightly in contact in pairs. Oral and adorai shields rather stout. Four or five oral papilla? on either side of each jaw ; the outermost one is pointed inwards above the next papilla, which is the largest. Teeth arranged in a single vertical row. Dental papillae absent. Genital slits not reaching to the disk margin. Arms short, very stout at base, rather rapidly tapered outwards to very slender extremities. Dorsal, as well as ventral, arm plates more or less rhomboidal, not very widely in contact or mostly separated from one another. Two to four arm spines, short, lying flat on the arm. One or two tentacle scales to each pore. This genus includes (1) species with two tentacle scales to each pore, viz., Ophiozona nivea Lyman, 1875, 0. tessellata Lyman, 1878, 0. marmorea Lyî^ian, 1883, 0. dypeata Lyman, 1883, 0. hlspinosa Kœhleb, 1897, 0. molesta Kœhlek, 1904, 0. elevata Clark, 1911, 0. platydisca Claek, 1911 ; (2) species with one tentacle scale to each pore, viz., 0. insuJaria Lyman, 1868, 0. stellata Lyman, 1878, 0. antillarum Lyman, 1878, 0. depressa Lyman, 1878, 0. alba Lütken & Moktensen, 1899, 0. contigua LtJTKEN & MoETENSEN, 1899, 0. casttt Kœhler, 1904, 0. depressa var. media Kœhler, 1904, 0. projecta Kœhler, 1905, 0. longispina Clark, 1908, 0. pokjplax Chxmi, 1911, and 0. (/*«//?wirt Mortensen, 1913, the genotype being 0. loiir/ispina. Key to Japanese species of Ophioxonella. A — Two tentacle scales to each pore. MONOGEAPH OF JAPANESE OPHIUROIDEA. 295 a — Disk convex, covered with one hundred to one hundred and fifty large plates elevata. act — Disk flat, covered with about two hundred small plates. . platydisca . A A — One tentacle scale to each pore, often also with one to three small, supx^lementary ones on the adradial border of the pore in a few basal arm joints. b — Disk convex ; arms long and slender, five to eight times as long as the disk diameter ; colour reddish or yellowish brown in alcohol. . . projecta. W>— Disk flat; arms short, tlu:ee or four times as long as the disk diameter ; colour white in alcohol. c — Disk plates fine ; radial shields rather small, about one-third as long as the disk radius ; arm spines shorter than the corresponding arm joint polypJax. cc — Disk plates large ; radial shields very stout, about half as long as the disk radius ; arm spines of basal arm joints longer than the corresponding joint. longispina, Ophio^ouella elevata (Clark). Ox>Mozona elevata : Clark, Biül. Ü. S. Nat. Mus., LXXY, 1911, p. 31, fig. 6. Off Goto Is., Eastern Sea; 95-100 fathoms (Claek). Ophio^onella platydisca (Clark). OpUozona plaUjdisca : Clark, Bull. U.S. Nat. Mus., LXXV, 1911, p 34, fig. 7. Off Hiro Misaki; 191 fathoms (Claek). Oj^hio^onella projecta (Kœhler). OpUozona j^rojecta : Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 19,. PI. T, figs. 16-18; Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 36. 296 ART. 2. — H. MATSUMOTO Numerous specimens ; off Nii-jima, Sagami Sea. Numerous specimens ; off Misaki, Sagami Sea. Off Goto Is., Eastern Sea ; 95-106 fathoms (Clark). Malaysian waters. Sagami Sea is pro- bably the northern limit of this Indo-Pacific species. The arms are very long and slender, being five to eight times as long as the disk diam- eter, so that they are longer than in Fig. 79. OpUozondla projeda. x 14. «. From above. J\.ŒHLEE S type. Ihô 5. From below, c. Side view of three arm joints near -^^^ ^^^^^^ ^^^^ ^^^ (lisk. ^ to each slit, the outer one being much longer than the inner. In some of the basal free arm joints, there are three arm spines instead of two. Ophiozonella j^olyplaoc (Clauk). Opliiozona polyplax : Clark, Bull. U. S. Nat. Mas., LXXV, 1911, p. 35, fig. 8. Off southern Japan ; 437 fathoms (Clark). Off Shio Misaki ; 440-587 fathoms (Cl.\rk). >rONOGiL\PH OF JAPANESE OPHIUROIDEA. 297 Ophioxonella longispina (Claiik). OpJiiokona longispina : Clakk, Bull. Mns. Comp. Zool., LT, 1908, p. 290; Clark, Bull. U. S. Nat. Mus., LXXY, 1911, p. 33. Numerous specimens ; Suruga Gulf. Numerous specimens ; Sagami Sea. Off Heda ; 337 fathoms (Clark). Suruga Gulf ; 45-124 fathoms (Clark). Uraga Channel ; 58 fathoms (Clark). Off eastern Japan ; 101 fathoms (Clark). The supplementary tentacle scales are very well developed. In the basal arm joints, there are one or two of them on the adradial border of each pore, beside one large, oval or- dinary scale on the abradial border. So that, the tentacle scales of this species, as well as of Ophio- ■lonella alba (Lütken & Mortensen), re- mind us of those of Ophiodoris and of certain species of Ophiochiton. The dorsal and ventral arm plates are re- latively wider than in many other species of this genus. The Fig. 80. Ophiozonella longispina. x3.3. a. From fibove. 5. From below, c. Side view of three arm joints near disk. 298 ART. 2. — H. MATSUMOTO : arm spines are also very long, in contrast to those of other species. Ophiozona (Lyman, 1865) restr. Disk covered witli very numerous small plates, the larger ones being surrounded by a zone of smaller ones. Kadial shields small, separated from one another. A trio of disk plates dis- tinctly present between the outer parts of each pair of radial shields. Two long genital slits in each interradius. Oral and adorai shields small. Four or five oral papilla3 on either side of a jaw ; the outermost one is pointed inwards above tlie next papilla, which is the largest. Teeth arranged in a single vertical row. Dental papilla3 absent. Arms rather long and slender, very gradually tapering outwards, with obtuse end. Dorsal as well as ventral arm plates well developed, almost quadrangular, widely in contact with one another through almost the entire length of the arm. Lateral arm plates of either side not meeting above or below. Four or five short arm spines. Two tentacle scales to each pore. This genus, as here restricted, includes Ophiolepis impressa LüTKEN, 1859, and 0. pacifica Lütken, 1859, the former being the genotype. Ophiozona is very near to Ophiolepis, being distinguished from it only by the total absence of supplementary dorsal arm plates. Ophiosona is wholly a littoral form, like Ophiolepis, and ranges to the West Indies and the Pacific side of Panama. The fauna^- of either side of Panama closely resemble each other, apparently as a result of a former open communication. The distribution of Ophiozona, as well as of Ophioderma, is interesting in this respect. JlONCGRArH OF JAPANESE OPHIUEOIDEA. 299 Key to Japanese species of Ophiolepls. A — Eadial shields about as large as the ordinary disk plates ; arms slender ; supplementary dorsal arm plates present along the outer and lateral borders of the dorsal arm plates proper ; three or four arm spines cincla. A A — Radial -shields much larger than the ordinary disk plates; arms stout ; supplementary dorsal arm plates present only ou either side of the dorsal arm plates proper ; five to seven arm spines , annulosa. OpMolextis cincta Müller & Troschel. OpJiiolepis cincta : Müller & Troschel, Sys. Ast., 1842, p. 90 ; Lüt- KEX, Adit, ad Hist. Oph., n, 1859, p. 101, PI. n, figs. 6a & 6b ; von Martens, Arcli. Naturg., XXXYI, 1870, p. 245'^; Lyman, Kep. Challenger, V, 1882, p. 19, PI. XXXVII, figs. 7-9 ; Studee, Abh. K. Preuss. Akad. Wiss. Berliu, 1882, p. 7 ; Marktanner-Tdeneretschee, Ann. K. K. Naturh. Hofmus., n, 1887, p. 294; Duncan, Journ. Linn. Soc. London, XXI, 1887, p. 8G ; Brock, Zeitschr. wiss. Zool., XL VII, 1888, p. 475 ; Bell, Proc. Zool. Soc. London, 1888, p. 388 ; Döderlein, Zool. Jahrb. Sys., m, 1888, p. 831 ; Loeiol, Piev. Suisse Zool., I, 1893, p. 398 ; Loriol, Mem. Soc. Phys. d'Hist. Nat. Genève, XXXH, 1894, p. 9; Kœhler, BuU. Sei. Fr. Belg., XXXI, 1898, p. 66 ; Ludwig, Abh. Senckenberg. Naturf. Gesell., XXI, 1899, p. 514; Pfeffer, ibid., XXV, 1900, p. 83; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 16; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, p. 287; Clark, Bull. Mus. Comp. Zool., LI, 1908, p. 289. OpTdolepis garrettii : Lyman, Piuc. Boston Soc. Nat. Hist., VHI, 1862, p. 77'^ Lyman, III. Cat. Mu^ Comp. Zool., I, 1865, p. 61, PI. H, fig. 4. Ophiolepis cincta vnr. nigra : Studer, loc. cit. Two specimens ; locality unknown. 1) These papers were not seen by me. 300 ART. 2. — H. MATSUMOTO Indo-Pacific. The disk is very convex and nearly lie- mispherical. In one of the specimens, the disk is 14 mm. in diameter and 6 mm. in height. The arms are of unequal length, and are two and a half times to five times as long as the disk diameter. The colour is mottled on the disk, and armulated on the arms, with greenish and yellowish gray. Fig. 81. Ophlolepis cinctn. x6. a. From above, h. From below. OX)hiole±}is amiiilosa Müller & Troschel.'^ Oplimra anmdosa : Blatnyille Maun. Act., 1834, p. 244, PI. XXIV, figs. l-4'>. (Non Lamarck, 1816.) OpJiioIepis annulosa : Müller & Troschel, Arch. Naturg., VI, 1840, p. 328-'; Müller & Troschel, Sys. Ast., 1842, p. 89, PL VIII, fig. 4 ; Lütken, Addit. ad Hist. Oph., H, 1859, p. 101, PI. II, figs. 5a & 5b ; Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, p. 58 ; yon Martens, Arch. Natnrg., XXXVI, 1870, p. 246-); Lyman, Rep. Challenger, V, 1882, p. 19; Bell, Proc. Zool. 8oc. Loudon, 1887, p. 140; Bell, ibid., 1888, p. 388; Döderlein, Zool. Jahrb. Sys., m, 1888, p. 831; Loriol, Mém. Soc. Pliys. d'Hist. Nat. Genève XXXII, 1895, p. 10; Bell, Proc. Zool. Sor^. London, 1894, p. 395 ; Döderlein, Semon - Zool. Forschuugs]-, V, 1896, p. 283 ; Ludwig, Abh. 1) Clark has recently proposed !i new name Ophlolepis siiperba for this species (Spolia Zeylanica, X, 1915, p, 80). Though Biainyille's Ophiura annulosa is preöccupied by La-maeck's, MuLLEB & Tkoschel's Oplùolepis annulosa is entirely free of homonyms. 2) These papers were not seen by me. MONOGEAPH OF JAPANESE OPHIUKOIDEA. 301 Senekeuberg. Naturf. Gesell., XXI, 1899, p. 544 ; Pfeffer, ibid., XXV, 1900, p. 83; Kœhlee, Exp. Siboga, XLI, Pt. 2, 1905, p. 17; Clark, Bull. Mus. Comp. Zool., LI, 1908, p. 289. One specimen ; Okinawa. Indo-Pacific. Okinawa is probably the northern limit of this species. The radial shields are larger and the space between each radial pair is narrower than in Lütken's figure. The colour is quite similar to Lütken's. The arm spines are only five or rarely six in number, so that they are fewer than in MüLLEE & Tkoschel's type. Fig. 82. Ophiolepis annulosa. x3. a. From above, h. Side view of three iirm joints neur disk. Key to Japanese species of Oplüoplociis, A — Disk scales more or less couvex, so that the surface of the disk is rather rough ; genital slits very short, pore-like ; the halves of each dorsal arm plate proper separated from each other by six larger and a few smaller supplementary plates imhricahts. AA — Disk scales flat, so that the surface of the disk is very smooth ; genital slits rather long, slit-like; the halves of each dorsal arm plate proper separated from each other by eight to ten larger and several smaller supplementary plates japonicus. Oxtliioplocus imhricatus (Müller k Troschel). Ophiolepis imhricata : Müller & Troschel, Sys. Ast., 1842, p. 93 ; LÜTKEN, Addit. ad Hist. Oph,, n, 1859, p. 160. OpMoploczis iessellatus : Lyman, Proc. Boston Soc. Nat. Hist., VIII, 1861, p. 76^>. 302 ART. 2. — H. MATSUMOTO : Ophioplocus imbricatus: Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, 13. 69 ; VON Martens, Arch. Naturg., XXXVI, 1870, p. 246^^ ; Lyman, Rep. Challenger, V, 1882, p. 20; Studer, Abh. K. Prenss. Akacl. Wiss. Berlin, 1882, p. 7 ; Marktanner-Tukneretscher, Ann. K. K. Naturh. Hofmus. Wien, ir, 1887, p. 295 ; Brock, Zeitschr. wiss. Zool., XLVII, 1888, p. 477 ; Bell, Proc. Zool. Soc. London, 1888, p. 388; Döderlein, Zool. Jahrb. Sys., Ill, 1888, p. 840 ; Loriol, Rev. Suisse Zool., I, 1893, p. 298 ; Loriol, Mém. Soc. Phys. cl'Hist. Nat. Genève, XXXII, 1894, p. 12 ; Pfeffer, Jahrb. Wiss. Anstalt, Hamburg, Xm, p. 47 ; Döderlein, Semon — Zool. Forschungsr., V, 1896, p. 283; Ludwig, Abh. Seuckenberg. Naturf. Gesell., XXI, 1899, p. 544; Kœhler, Exp. Siboga, XLY, Pt. 2, 1905, p. 333; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, p. 288 ; Clark, Bull. Mus. Comp. Zool., LI, 1908, p. 289 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 30. One specimen ; Okinawa. Indo-Pacific. Okinawa may be the northern limit of this species. Opliioiylocus Japonicus Clark. Ophioplocus imbricatus : Lyman, Bull. Mus. Comp. Zool., Ill, 1874, p. 228 (pars); Marktanner- Turneretscher, Ann. K.K. Naturh. Hofmus., 11, 1887, p. 295 ; Ives, Proc. Acad. Nat. Sei. Philadelphia, 1891, p. 4, PI. V, figs. 6-10. OpJdoplocus japonicus : Clark, Bull. Ü. S. Nat. Mus., LXXV, 1911, p. 30, fig. 5. Namerous specimens ; Eno-m^a, Surnga. Numerous specimens ; Misaki. Numerous speci- mens ; Kominato, Bôshù. Pacific coast of Honshu, probably eastwards to the limit of the Kuroshiwo Sti-eam. Fig 83. Ophioplocus Imhricutus. x 5. a. Side view of three firm joints near disk. b. Dorsal view of four arm joints near disk. 1) These papers were not seen by me. MONOGRArH OF JAPANESE OPHIUEOIDEA. 303 This species is distinguished from 0. imhricatus chiefly by tlie longer genital slits and by the more numerous supplementary dorsal arm plates. Further, the disk scales and the supplementary plates are coarser, thicker and more convex in 0. imhri- catus than in 0. japoniciis. I suppose that, Maektanner-Tukneeetschee's specimens from Enoshima are to be referred to the present species, because the occurrence of 0. imhricatus there is very improbable. 0. japonlcus is undoubtedly the most common ophiuran around Misaki, and perhaps along the entire southern coast of Honshu. It occurs together with Ophiarachnella gorgonia, but is far more abund- ant. In a large specimen in my hands, there are four arm spines in some of the proximal free arm joints. Fig. 84. Ophioplocus japonicus. x 5. a. Side view of three .arm joints ne.ar disk. h. Dorsal view of four arm joints near disk. Family 2. Ophioleucidse Matsumoto, 1915. Disk flat, covered with very thick scales with superficial granulations. Radial shields naked or covered with granules ; in internal view, they are very large and joined or very close -set in pairs. The radial shield and genital plate articulate with each other by means of two condyles and one pit. Genital shts very long, extending from the oral plate to the disk border, sometimes onto the dorsal side of the disk. Oral and adorai shields of moderate size or often large ; each interradial pair of the latter joined to each other just inside the former. Numerous close -set oral papillae, arranged in a single row. Dental papillae absent. Teeth arranged in a single vertical series. Peristoraial plates 304 AKT. 2. — H. MATSUMOTO : usually double or triple, of moderate length or sometimes very long. The genital plate and scale of the same side of a radius articulate with each other near the outer end, both being very long and narrow. Arms very long and slender, inserted ventrally to the disk. Arm plates all well developed, or sometimes the dorsal and ventral ones may be very rudimentary. Two, or sometimes more, arm spines, short, conical, lying flat on the arm. One or two tentacle scales to each pore, which is sometimes very large. This family includes seven genera, which may be grouped as follows. I. Single tentacle scale ; dorsal arm plates well developed, widely in contact with one another. Ophiopcepale Ljungman, 1871. OpMocirce Kœhlee, 1904. Ophioleuce Kœhlee, 1904. OphîopaUas Kœhlee, 1904. IL Single tentacle scale ; dorsal arm plates very rudimentary, extremely minute. Ophiotrochus Lyman, 1878. III. Two tentacle scales ; dorsal arm plates well developed, widely in contact with one another. OpJiiernus Lyman, 1878. Ophiopyren Lyivlin, 1878. In Ophioleuce charischema (Claek), the peristomial plates are moderately large, triple, the paired ones being usually separated from each other by a median one. The oral plates are long and very slender. The oral fi-ames are also long and slender, without lateral wings. The basal vertebra^ are relatively very long, the first one being the shortest and more or less discoidal. The genital plates are long and bar-like, with two articular condyles MONOGRAPH OF JAPANESE OPHIUEOIDEA. 305 and one articular pit at the enter end to fit to the t^vo condyles and one pit of the radial shield. The genital scales are long, wide, thin, leaf-hke, L-shaped in transverse section, articulating with the genital plate near the outer end of the latter. The radial shields in internal view are very large and completely joined in pairs. Ophiojjœpale goèsiana Ljungman and Ophiermis valUncola Lïman, which were dissected by Lyman, appear to be essentially similar to the preceding in the internal structures, save that the peristomial plates of the first are not triple but double, and those of the second single for each oral angle. Again, according to Lyman, the peristomial plates of Ophiopyren Jongispiniis liiiviAN are double and very long, forming together a complete ring around the mouth. As a wiiole, the internal structures of the present family resemble those of OpMura and its allies. Key to Japanese genera of Ophioleiicidœ, A — Tentacle pores small^ each provided with a single scale. a — Oral papillse not soldered together, uniform in size, or the outer- most one is larger than the others ; arm joints short ; dorsal arm plates well developed, widely in contact with one another ; arm spines uniform in length Ophioleiice. au — Oral papillœ soldered together, except the outermost two, which are exceedingly large ; dorsal arm plates very rudimentär}', exceedingly minute ; basal arm spines especially long, the others short OpMotrocIms. ^^— Tentacle pores large, each provided with two scales ; dorsal arm plates very well developed ; ventral plates entire, in contact with one another ; arm spines uniform in length Ophiernus. 306 ART. 2. — H. MATSUMOTO: Key to Japanese species of Oi?hioleuce. yi— Dorsal side of disk, including the radial shields, closely covered with granules ; oral shields longer than wide ; outermost oral papilla not especial!}' large; arm spines more than half as long as the corresponding arm joint char i schema. AA — Eadial shields, as well as the larger disk scales, paiiially free of granules ; oinl shields wider than long ; outermost oral j)apilla very large and operculiform ; arm spines less than half as long as the corresponding arm joint hrevispinum. _„;:/l!;^!-jj. Ophioleuce charischema (Of^akk). Ophiocten charischema : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 97, fig. 35. Numerous specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Eastern Sea; 95-152 fathoms (Claek). Off Honshii ; 63 fathoms (Claek). In my specimens, several spines are present on the disk, mingled with granules ; they fall off easily on rubbing. The smaller specimens have the dorsal disk plates and radial shields partially naked. Ophioleuce hrevispinum (Clark). Fig. 85. Ophioleuce charischema xO. a. From above, h. From below, c. Side view of four arm joints near disk. Ophiocten hrevispinum : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 98, fig. 86. MoNOGRAPH OF JAPANESE OPHIUROIDEA. 307 Eastern Sea ; 301 fathoms (Clark), OpMotrocUiis lotigispinus Clap.k. OpUotrochm lonrjisplnm : Cr.APJv, Bull. U.S. Nat. Mus., LXXV, 1911, p. 94, fig. 33. Sagami Sea ; 918 fathoms (Clark). Ophiernus adspersus LY.AfAN. Ophlernus adspersus : Ly.aüan, Bull. Mus. Comp. Zool., X, 1883, p. 236, PI. Ill, figs. 19-21; Kœhlee, Anu. Sei. Nat. Zool., Sér. 8, IV, 1896, p. 316 ; Kœhler, Ecli. Indian Mus., Deep-sea Opli., 1899, p. 32 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 96, fig. 34. Ophkrnus annedem : Lütken & Moktensen, Mem. Mus. Comp. Zool., XXIII, 1899, p. 107, PI. V, figs. 4-6. Eastern Sea ; 440 fathoms (Clark). Colnett Strait ; 1,008 fathoms (Clark). Off Hiuga ; 578 fathoms (Clark). Off Kii ; 244-290 fathoms (Clark). West Indies. Eastern Atlantic. Eastern and western Pacific. Indian Ocean. Family 3. Ophiodermatidae Ljuxgman, 1867. Disk closely covered with fine grannies, and sometimes with scattered spines. Oral angles, sometimes also the oral shields, covered with granules. Numerous close-set oral papilla?, of wliich the outermost is pointed inwards and projects above the next papilla, which is the largest. Dental papilla) absent. Peristomial plates triple. Oral frames without well developed lateral wings. Ptadial shield and genital plate of the same side of a radius arti- culating with each other by means of two condyles and one pit. 308 ART. 2.— H. MATSÜMOTO: Arms moderately long, stout, cylindrical ; stoutest at the base, horizontally flexible. Arm plates all well developed. Numerous arm spines, short or moderately long, lying flat on the arm or erect. One or two tentacle scales to each pore ; when there are two, the abradial scale overlaps the base of the lowest arm spine. This family includes sixteen genera, which may be grouped into two subfamihes as follows. Subfamily 1. Ophiarachnînœ Matsumoto, 1915. — Arm spines not very short, erect. I. Oral shields entirely covered with granules ; arm plates usually concentrically striated ; arm spines hyaline. Ophiurocœns Matsumoto, 1915. Ophiurodon Matsumoto, 1915. II. Oral shields naked ; arm plates not concentrically striated ; arm spines opaque. Ophiurochceta Matsumoto, 1915. OpMarachna Müllee t^- Tkoschel, 1842. Subfamily 2. Ophiodermatinœ Matsumoto, 1915. — Arm spines very short, lying flat on the arm. A. Single tentacle scale to each pore. I. Oral shields entirely covered with granules. Ophioconis LiJTKEN, 1869 { = Ophiocormus Clakk, 1915).^^ 1) I can not find in Clark's description and figures any tangible character, by which his OpJiioconmcs is distinguished from the genuine OpTdoconis. The bhmt tip of the arms and the very short and apprcssed arm spines of his Ophiocormus are not at all Ophiocanthine but perfectly Ophiodermatine. Clakk's unique specimen, on -which his genias is based, is certainly very young, measriring only 3 mm. across the disk and 6 mm. in the arm length. So that, all the " notable " characters are merely what may be expected in such a young specimen. In my opinion, Ophiocormus is a genuine Ophioconis, if indeed it is not a yoiing stage of CrifftoixKa. MONOGRAPH OF JAPANESE OPIIIUliOIDEA. 309 Cryptopelia Clakk, 1909. II. Oral sliields naked. Bathypectinura Clark, 1909. B. Two tentacle scales to each pore. III. Genital slits entire. Pectinura Forbes, 1842. Ophiopezella Ljungman, 1871. Ophiochce'a Lütken, 1869. Ophiarachnella Ljungman, 1871. Ophiochasma Ljungman, 1871- IV. Genital slits divided into two secondary pores. Ophioderma Müller & Troschel, 1842. Ophioncus Ive«, 1889. Diopederma Clark, 1912. Ophiocryptus Clark, 1915. This family is very uniform in its internal structure. The peristomial plates are always triple, one of the three secondary plates filling up the outer open angle formed by the other two, as observed by myself in Ophiuroconis monolepis Matsumoto, Ophmrodon grandisquama (Kœhler), Ophîurochœta mixta (Lyman), Ophiarachna mcrassata (Lamarck), Pectinura anchista Clark, Bathypectinura goto I Matsumoto, Ophiarachnella gorgonia (Müller & Troschel), Ophio- derma januarii Lütken, 0. brevicauda Lütken, &c. In Ophiarachia ' incrassata and Bathypectinura gotoi, the oral frames are very long and have distinct, V-shaped grooves for the ambulacral ring canal, while in the others they are short and have no distinct grooves for the ring canal. In Ophioderma, the genital plate and scales of the same side of a radius are soldered together at the middle, so that the genital slit is divided into two secondary pores, an inner and an outer. An important characteristic of the present 310 ART. 2.— H. MATSUMOTO : family is that the radial shield and genital plate of the same side of a radius articulate with each other always by means of two articular condyles and one articular pit. Revision of Ophioconis, s. ext. In my opinion, Ophioconis as hitherto understood is a very heterogeneous genus and is divisible into many natural groups, each of which is quite compact and worthy of forming a distinct genus. In treating of Ophioconis diastata and papillata, Claek^^ expresses a serious doubt as to the systematic position of Ophio- conis s. ext., and remarks that, he will " not be surprised if some of the species now placed in Ophioconis, really belong in the Ophiacanthidce, while others are placed in the Ophiodermatidœ.'" Kœhler'' also says that, " les 0. cupida, permixta, cincta et grandisquama forment, dans le genre Ophioconis, un groupe à part et qui offre une physionomie différente de celle des autres Ophio- conis y Both authors are right in their views. Claek's species are, in my opinion, referable to Ophiolimna, while the species mentioned above by Kœhler as forming a separate group are to be included in Ophiurodon Matsumoto. A third atypical group of Ophioconis is represented by 0. miliaria Lyman, 1878, and j^wZver?^- lenta Lymak, 1879, which I wish to refer to Ophiuroconis Matsu- moto, while a fourth is represented by Ophioconis indica Kœhlee, 1898, for which however I do not dare to propose a new generic name, because I have no specimens of it. Thus the genuine Ophioconis includes only two Mediterranean species, viz. 0. forhesii (Hellee, 1862) and hremspina Ludwig, 1880. The new sub- 1) Bull. U.S. Nat. Mus., LXXV, 1911, p. 28. 2) Exp. Siboga, XLV, Et. 2, 1905, p. IR. MONOGRAPH OF JAPANESE OPHIUROIDEA. Bll divisions of Ophioconis s. ext,, together with certain aUied genera, are distinguished as follows. A — Arm spines long and flagellate, erect. a — Outermost oral papilla very large and operculiform ; peristomial plates simple, verj short and wide, the two halves being soldered together fairly well, Avithout median secondary plate Ophiolimna, emend. aa— Outermost oral papilla pointed inwards, projecting above the next papilla, which is the largest; peristomial plates triple, with a median secondary plate, fairly long and wide. h — Oral shields entirely covered with granules ; arm plates usually concentrically striated ; arm spines hyaline, c — Teeth triangular and pointed, ventral arm plates wider than long, usually separated from one another OpMuroconis. cc — Teeth flat, thin, with much widened and often serrate end ; ventral arm plati s longer than wide, in contact Avith one another . ; .- Ophiiirodon. lb — Oral shields naked ; arm plates not concentrically striated ; arm spines opaque. d — Disk granulations mingled with a number of spines ; accessory oral shields absent Ophmrochœta . dd — Disk free of spines ; accessory oral shields usually present just out- side the ordinary ones Opidarachia . AA — Arm spines very shoi-t, lying flat on the arm, hyaline. e — Oral shields entirely covered A^ith granules Ophioconis, restr. ee — Oral shields naked '' Ophioconis " indica. OjyJiioiimna emend, is il typical nienaber of the Ophiacanthidce ; Ophiuroconi's, Ophiurodon, Ophiurochceta and Ophiarachna form the Ophiaraclmince ; and Ophioconis restr. and " OpJnoconis " indica are referable to the Ophiodei^mcdinœ. Oplnoconis restr. is very near to Cryptopelta, but differs from it in the hyaline arm spines, while " Ophioconis'' indica resembles Pectinura, differing from it also in 312 ABT. 2. — H. MATSUMOTO : the hyaline arm spines. There is some possibihty that, Ophioconis and Cryptopelta are congeneric and that " Ophioconis " iîidica is an extreme form of Pectinura. Ophiuroconis Matsumoto, 1915. Disk and oral angles, including the oral shields, closely covered with fine grannies. Six or seven oral papillae on either side of a jaw ; the outermost one is pointed inwards and projects above the next papilla, which is the largest. Teeth triangular and obtusely pointed. Dental papillae absent. Arms not very long, cylindrical, widest at the base, tapering outwards to a very slender tip, where the vertebrae are imperfectly divided into halves by a series of pores. Ventral arm plates wider than long, not in contact with one another except in the very basal joints. Six or more arm spines, long, flattened, hyaline, laterally flared. One or two tentacle scales to each pore. This genus includes Ophioconis puherulenta IjYMan, 1879, and 0. miliaria Lyman, 1878, besides the genotype, Ophiuroconis monolepis Matsumoto, 1915. OpJiiiirocoiiis inonolepis Matsumoto. OpMuroconis monolepis : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 85. Six specimens ; Sengendzuka-Aoyamadashi, Sagami Sea ; 85 fathoms. Two specimens ; Mera-out-Oisegaké, Sagami Sea ; 300 fathoms. Diameter of disk 5 mm. Length of arms 25 mm. Width of arms at base 1 mm. Disk nearly circular, slightly puffed interbrachially, very soft, MONOGRAPH OF J.iPANESE OPHIUEOIDEA. 313 closely covered with very fine granules, of which eleven or twelve lie in 1 ram. Radial shields entirely covered over. Interbrachial ventral surfaces and oral angles, including the oral and adorai shields, also covered with granules, which are similar to those of Fig. 8b. Ophmroconis monolepls. d. From above. xl5. h. From below. xl2. c. Side view of two arm joints near disk, x 12. the dorsal side. Genital slits long, nearly reaching to the disk margin. Six or seven oral papilte on either side, closely set, acute ; the outermost one is pointed inwards, projecting above the next papilla, which is slightly larger than the rest and obtuse. Teeth triangular, not very stout, obtusely pointed. Arms about five times as long as the disk diameter, more or less cylindrical, widest at the base, uniformly tapered outwards to a very slender tip. Dorsal arm plates rather small, fan-shaped, with convex outer border, slightly shorter than the corresponding arm joint, wider than long, convex along the median line, so that the arm is keeled as a whole. Lateral arm plates of the two sides meeting above as well as below. First ventral arm plate extreme- 314 ART. 2. — H. MATSTJMOTO : ly small and very insignificant. Those following small, triangular, with very large inner angle and very long, cont^ex outer border, shorter than the corresponding arm joints, much wider than long ; they become smaller and shorter outwards. All the arm plattes are concentrically striated, as seen under the microscope. Six or seven arm spines, spiniform, acute, slightly flattened, hyaline, erect ; the uppermost one or two are nearly twice as long as the corresponding arm joint, while the lowest one is slightly shorter than the same. Single small, leaf-like, acate tentacle scale to each pore. VertebrcX of the distal arm joints imperfectly divided into halves by a series of pores. Colour in alcohol light yellow. This species is at once distinguished from both 0. miliaria (Lyman) and pulverulenta (Lyman) by the fewer oral papilla?, by the fewer and shorter arm spines and by the presence of only a single tentacle scale to each pore. Ophiurodon Matsumoto, 1915. Disk, as well as radial shields, closely covered with fine granules, sometimes bearing scattered spines. Oral angles and oral shields also granulated. Four or five oral papillae on either side, the outermost one projecting inwards above the next papilla. Teeth very flat and widened, with very thin and translucent end, arranged in a single vertical series. Dental papillas absent. Arms not very long, widest at the base, tapering outwards to a very slender tip. Dorsal and lateral arm plates usually concentrically striated. Ventral arm plates very narrow, longer than wide, fully in contact with one another. Vertebra? of the distal arm joints often imperfectly divided into halves by a series of pores. More MONOGEAPH OF JAPANESE OPHIUROIDEA. 815 than six arm spines, long, flared, longer than the corresponding arm joint. Single tentacle scale to each pore. This genus includes Ophioconis cinda Bkock, 1888, 0. permixta Kœhlee, 1905, and 0. cupida Kœhlek, 1905, besides the genotype, 0. grandi squama Kœhlee, 1904. OpMurodon grandisqumna (Kœhlek). OpJiioconis gradisquama : Kœhlee, Exp. Siboga, XLY, Pt. 1, 1904, p. 11, PI. n, figs. 11 & 12. One specimen ; Okinosé, Sagami Sea. Malaysian waters. The teeth are very flat, thin, widened, hyaline, but not so sharply serrate as in the three other species of the present genus. Fig. 87. Opkmrodon grdndlsquama. a. From above. x20. h. From below, x 12. The dorsal and lateral arm plates are concentrically striated. The arm spines are longer than in the three other species, and are much flattened, not uniformly tapered, but retain nearly the same width until very near the end, where they abruptly taper. In the flrst few arm joints, they are more or less spatulate. 316 ART. 2. — H. MATSUMOTO : Ophiurochseta Matsümoto, 1915. Disk, as well aa radial shields, closely covered with fine grannies and bearing scattered spines. Oral angles granulated, oral shields naked. Oral papillae numerous, close-set, the outermost papilla being pointed inwards above the next papilla, which is the largest. Arms not very long, rather stout, widest at the base. Dorsal, as well as ventral, arm plates well developed, fnlly in contact with one another. Six or more arm spines, long, flared, longer than the corresponding arm joint. Two tentacle scales to each pore, the abradial one overlapping the base of the lowest arm spine. This genus includes Ophiochceta mixta Lyman, 1878 (referred to OphioUmna by Verkill, 1899), and Ophiolimna littoralis Kœhler, 1912, the first being the genotype. This genus apparently resembles Ophiochceta, bat differs from it in the long and fiared arm spines. Ophiurochœta differs from Ophiolimna Veerill, 1899, in the more numerous oral papillae, of which the outermost one is not very large and operculiform, but pointed inwards above the next papilla, which is the largest ; in the well developed dorsal and ventral arm plates, and in the pre- sence of two tentacle scales, of which the abradial one overlaps the base of the lowest arm spine. Verriijl- thinks that the internal structures of 0. mixta are much like those of Ophiacantha ; but my own observations lead to a quite different conclusion. I have observed the internal structures of Ophiuroconis mono- lepis, Ophiurodon grandisquama, Ophiurochœta mixta, Ophiarachna incrassata, Ophiochiton fastigatus, Ophioplax lamellosa, &g., and satisfied myself that they all belong to the same typo. In these forms, the peristomial plates are always triple, one of the secondary MONÜGRAPH OF JAPANESE OPHroßOlDEA. 3^17 plates filling up the outer open angle formed by the other two ; while in Ophiacantlia, OpliioUmna, &c., the peristomial plates are simple, or double with soldered halves, always lacking the unpaired secon- dary plate. Further, the peristomial plates are relatively to their width distinctly longer in the former type than in the latter. When I compared a young specimen of Ophiarachia incrassata with a specimen of Ophiurochceta mixta, kindly presented by Dr. H. L. Claek, and with Kœhler's photographs of 0. Uttoralis, I was instantly struck by the similarity in their plan of structure. The only differences of Ophiurochceta from Ophiarachna are the presence of the scattered disk spines and the absence of the accessory oral shields, besides the smaller size. The systematic value of the accessory oral shields is, however, scarcely recognised by Claek. And I have also observed the absence of accessory slaields in some interradii of a certain specimen of Ophiarachna incrassata. One may wâtli good reason regard the relation of Ophiurochceta to Ophiarachna to be parallel to that of Ophiomastix to Ophiocoma or of Ophiochœta to Pectinura.^^ 1) Claek, 1915, has i^rovisionally referred my OpMurochccta to Verbihl's Opfdotreta. The l.atler genus, which includes six species .iccording to Claek, is undoubtedly too heterogeneous to be looked upon as n distinct genus ; and I can find practically no common characteristic for these six species. I look upon Veeeill's Ophiotreta merely as a section of the genuine Ophiacantha. As Claek's Ophiotreta is A'ery heterogeneous, it is almost useless to compare my Ophiurochceta with the rest of Claek"s Ophiotreta. So that I wish here to compare my Ophiurocho'ta with the type of OphiotrePi. They differ as follows. OpMurochada : — Disk covered with fine granules, besides scattered spines ; oral angles regularly triangular, regularly closing the oral slits, distinctly granulated; oral shields situated far in- wards, so that the distance from their inner end to the tip of the oral angles is very short ; no cluster of dental ijapillfc, though a few infradental papiUœ may be present ; orsal arm plates fairly wide, so that the arm spines do not approximate dorsally ; ventral arm i)lates dis- tinctly in contact with one another ; arm spines not extremely long, opaque and not serrate (at least in adult) ; two tentacle scales lo each pore, the abradial one overlapping the base of the lowest arm spine. Ophiacantha [Ophiotreta) Uneolata : — Disk covered with coarse granxxles, besides scatteretl spines ; oral angles distinctly narrowed at base by the second oral tentacle pores, so that the oral slits are gaping ; no granulation on the oral angles ; oral shields situated far outwards from 318 ART. 2. — H. MATSUMOTO. Ophiarachna incrassata (Lamarck). Opldura incrassata : Lamarck, Hist. Nat. Anira. saus ^''ert., II, 1816, p. 542. Ophiarachna incrassata : Müller & Troschel, Sys, Ast., 1842, p. 104 ; LÜTKEN, Addit. Hist. Opli., Ill, 1869, p. 33 ; Lyman, Bull. Mus. Comp. ZooL, III, 1874, p. 221 ; Lyman, Kep. Challeuger, V, 1882, p. 173 ; Brock, Zeitschr. wiss. Zool., XLYH, 1888, p. 495 ; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 64 ; Kœhler, Bull. Sc. Fr. Belg., XLI, 1907, p. 330 ; Clark, Bull. Mus. Comp. Zool, LI, 1908, p. 298; Clark, ibid., LH, 1909, p. 128. Three specimens ; Bonin Is. One specimen ; Okinawa. Indo-Pacific. Okinawa and Bonin Is. seem to be the northern limits of this species. One of the specimens from the Bonin Is. and the one from Okinawa agree perfectly with Müller à Tkoschel's description. The largest one from the Bonin Is. is 50 mm. in the disk diameter, 170 mm. in the arm length and 8 mm. in the arm width at base. Its colour in alcohol differs from that of the others, and is yellow- ish brown above, dark brown below, being handsomely dotted with lighter shade on the ventral interbrachial spaces, on the oral shields, on the tentacle scales, and on the lower arm spines. The tliird specimen from the Bonin Is. is notably different from the others : the oral shields are a little lonsrer than wide, and the the tip of the oral angles ; cluster of dental iJapillœ j)resent ; dorsal arm plates very narrow, so that the arm spines more or less approximate dorsaUy in a few basal arm joints; ventral arm plates entirely seiDarated from one another ; arm spines extremely long, translucent and finely serrate ; single tentacle scale to each pore (two scales to the first pore only), not overlajiping the base of the lowest arm spine. All the distinctive characters of Ophiurochœta are perfectly Ophiodermatine, and those of the type of OpJdoireta thoroughly Oi>hiaeanthine. I believe that, Ophiurochœta cannot be united \\ith Ophiotreta, unless the majority «jf the now recognised genera of the Ophiuroidca should be united to a single gentis. MONOGRAPH OF JAPANESE OPHIUROIDEA. 319 Fig. 8S. Ophifirachna incrassatn. xl^. '(. From above, h. From below, c. Skle view of four arm joints near disk, showing the regenerated arm spines. accessory oral shields are very rudimentary or absent.'^ Compar- ing these fom- specimens, we find that the arm spines are longer, blunter and more flattened in the larger specimens. The propor- tion of width to length of the oral shields is not constant. The size ratio of the accessory oral shield to the oral shield proper differs in specimens. The pairs of the adorai shields are sometimes asymmetrical. The arm spines may regenerate. Key to Japanese genera of Ophioderniatinœ. y^^ Single oval tentacle scale to each pore ; three or four arm spines ; radial shields partly naked Bathypectinura. ./i^— Two tentacle scales, of which the abradial one overlaps the base of the loAvest arm spine ; six or more arm spines. a — Radial shields covered over Pectinura. aa — Radial shields naked OpJdaraclinélla. 1) lu Pectinura, Ophiarachnella, Bathypectinura and Ophiarachna, the accessory oral shields are not constantly present. I therefore consider it more natural to refer Ophiopeza dnnhyi FABQrHAR to Ophiarachna than to I'ect'mnra, notwithstanding the absence of the accessory oral shields. 320 ART. 2. — H. MATSUMOTO Bathypectinura gotoi Matsumoto. Bathypedimira gotoi : Matsumoto, Proc. Acad. Nat. Sei. Philadelphia, 1915, p. 87. Two specimens ; Nishinoyodomi, Sagami Sea ; 170 fathoms. Diameter of disk 50 mm. Length of arms 195 mm. Width of arms at base 7 mm. Disk pentagonal, flat, closely covered with fine granules, of which fom' or five occur in 1 mm. Radial shields only partly naked, but distinguishable ■■:" through the superficial ^ . ■ • granulations as large, elon- gated ovate, slight swel- lings, nearly half as long as the disk radius, and wider outwards ; the naked part is very small, ovate, and wider without than within. Genital slits very long, almost reaching to the margin of the disk. Genital plates distinct, long and very stout. Oral shields small, triangular, with rounded angles and convex sides, almost as wide as long. Accessory oral shields very rudimentarv. In one of the two specimens, they Fig. 89. Bathypectinura rjoto'i. xl}. a. From above, h. From below, c. Side view of four arm joints near disk. d. Dorsal view of five arm joints somewhat near disk, showing the divided dorsal arm i:)lates. MONOGliAPH or JAPANESE OPHIUKOIDEA. 321 are absent ; but in tlie other tliey are indistinctly represented by one or two small scales, separated from the oral shield by gran- ules. The adorai shields are almost, and the oral plates entirely, covered with granules, which are coarser and sparser than in the outer parts. Eight or nine oral papilla? on either side ; the outer- most two or three are large, flat, thin ; the second from the outermost is the largest ; inner ones very small, somewhat conical, obtuse. At the apex of the jaw, an unpaired infradental papilla is often present. Five to seven teeth, irregular in shape and size, with pointed or rounded ends, arranged in an irregular vertical row. Arms long, stout, gradually tapered, with a rather sharp dorsal ridge, triangular in transverse section. Dorsal arm plates large, almost occupying the entire dorsal surface of the arm, quadrangular, with rounded outer corners, a little wider without than wdthin, much wider than long, three to four times as wide as long, with a rather sharp ridge in the median line ; some of them divided into several irregular pieces. Lateral arm plates very low, less than half the height of the arm, meeting neither above nor below. Ventral arm plates small, rhomboidal, with the shorter diagonal parallel to the arm axis. The first plate is almost as wide as, but much shorter than, the following. The three or four plates beginning with the second have a median keel, which is more prominent proximally both with regard to one plate and to the arm as a whole. Arm spines four for the most part, but three in the distal part, very short, flattened, lanceolate, obtuse ; the lowest one is somewhat longer than the others. One tentacle scale, large, oval, thin, flat ; adorai ones larger. Colour in alcohol light yellowish brown. This species is quite near to B. Jacertosa (Lyman), but 322 ART. 2. — H. MATSUMOTO .* differs from it in the coarser disk granules, in the smaller naked part of the radial shields, in the more strongly ridged dorsal arm plates, in the "much lower lateral arm plates, in the ventral arm plates being wider than long and angled inwards as well as outwards, and in the arm spines being much shorter than the corresponding arm joints. I*ectinura anchista Clark. Fectinura anchista : Clark, Bull. U. S. Nat. Mus,, LXXV, 191 1, p. 23, fig. 1. Numerous specimens ; off Misaki, Sagami Sea. One specimen ; off Uki-shima ; 300 fathoms. Off Suno Saki, Sagami Sea ; 49 fathoms (Clark). Kagoshima Bay ; 85 fathoms (Clark). Eastern Sea ; 95-139 fathoms (Clark). This species strongly resembles P. cylindrica (Hutton), and it is somewhat doubtful whether the two are really distinct. As to the proportion of the arm length to the disk diameter, the dif- ference pointed out by Clark has, in my opinion, no significance ; for the specimens in my hands have proportionately shorter arms. One specimen is 9 mm, in the disk diameter and 30 mm, in the arm length, while another specimen is 8 mm. in the disk diameter and 33 mm. in the arm length. The colour seems to change in alcohol with time from purple to pink. Key to Japanese species of Ojyfiiarachnella. A — No trio of naked disk plates just outside and between each pair of radial shields ; oral shields longer than wide yorgonwt. AA — A trio of naked disk plates present just outside and between each pair of radial shiekls, on the dorsal side of the arm base. MONOciRAPH OF JAPANESE OPHIUROIDEA. 323 a — Radial shields small, uot much larger than one of the trio of naked plates ; oral shields wider than long infernalis. aa — Radial shields very large, but the trio of naked plates very small ; oral shields longer than wide megaîaspis. Ox)hiavachnella (jorgonia (Müli>er & Thoschef;). OpJiiarachna gorgonia : Müller & Troschel, Sys. Ast., 1842, p. 105; Lyman, Rl. Cat. Mus. Comp. ZooL, I, 1865, p. 39. Pedimira gorgonia : Lütken, Addit. Hist. Oph., HI, 1869, p. 15 ; LYifAN, Rep. Challenger, V, 1882, p. 15; BEr.r., Rep. Zool. Coll. Alert., 1884, p. 134; Brock, Zeitschr. wiss. Zool., XLMI, 1888, p. 471; Doder- LEIN, Zool. Jahrb. Sys., Ill, 1888, p. 880; Loriol, Rev. Suisse Zool., I, 1893, p. 397 ; Kœhler, Ech. Indian Mus., Shallow-wat. Oph., 1900, PI. I, figs. 1 & 2; Pfeffer, Abh. Senckenberg. Naturf. Gesell, XXV, 1900, p. 83; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 8 ; Kœhf.er, Bull. Sei. Fr. Belg., XLI, 1907, p. 284; Clark, Bull. Mus. Comp. Zool., LI, 1908, p. 289. Pedinura marmorata : Lyman, Bull. Mus. Comp. Zool., Ill, 1874, p. 222, PI. V, figs. 1-7 ; Lyman, loc. cit., 1882, p. 17. Pedimira ramsayi : Bell, Proc. Zool. Soc. London, 1888, p. 281. Pedinura intermedia : Bell, loc. cit., 1888, p. 386. Pectinura stearnsii : Ives, Proc. Acad. Nat. Sei. Philadelphia, 1891, p. 212, PI. Xr, figs. 1-5. Pedinura venusta : Loriol, Mém. Soc. Phys. d'Hist. Nat. Genève, XXXII, p. 16, PI. XXIII, figs. 3-3h. OpJiiaracJineUa gorgonia : Clark, Bull. Mus. Comp. Zool., LII, 1909, p. 117. One specimen (belonging to the Seventh High School, Kago- sliima); Kagoshima Bay. One specimen ; Eno-m-a, Sumga. Numerous specimens ; Arai Beach, Misaki. This Indo-Pacific species is one of the most common ophiurans in the vicinity of the Misaki Marine Biological Station. The first specimen is very typical, and has a coarse squama- 524 AET. 2.— H. MATSUMOTO : tion distinguishable throngli the superficial granulations. The rest are of the steamsii-typG, being different from the typical gorgonia- type in the disk scales being fine, flat and almost indistinguishable through the superficial granulations. The specimen from Eno-ura, however, rather resembles the venusta-tj'pe in the shape of the oral shields and accessory oral shields. The largest specimen is 26 mm. in the disk diameter and 95 mm. in the arm length. The number of the arm spines increases with the growth of the animal. Ophiaraetmella infemalis (Müller & Troschel). OpUaraclma infemalis: Müller k, Troschel, Sys, Ast., 1842, -ç. 105. Fectinura in/erncdis : LÜTKEN, Addit, Hist, Oph,, III, 1869, p. 33 ; Lyman, Bull. Mns. Comp. Zool., HI, 1874, p. 222, PI. YII, fig. 1 ; Lyman, Hep. Challenger, V, 1882, p. 17 ; Bell, Rep. Zool. Coll. Alert, 1884, p. 134, PI. VIII, fig. B; Brock, Zeitsclir. wiss. Zool., XLVH, 1888, p. 471; LoEiOL, Bev. Suisse Zool., I, 1893, p. 397 ; Pfeffer, Abli. Senckenberg. Naturf. Gesell., XXV, 1900, p. 83; Kœhler, Exp. Biboga, XLV, Pt. 2; 1905, p. 7, PI. I, figs. 1-3; Kœhler, Bull. 8ci. Er. Belg., XLT, 1908, p. 285; Clark, Bull. Mus. Comp. Zool., LI, 1908, p. 289. Pecfinura similis : Kœhler, loc. cit., 1905, p. 6, PI. I, figs. 4-6. OphiarocImeUa infer- nalis : Clark, Bull. Mus. Comp. Zool., Ln, 1909, p. 124. Two specimens ; Okinawa. One speci- A men: Eno-shima, Osumi. Two specimens ; Encv ur;i, Suruga. Indo-Pacific. The radial shields Fig. 90. Ophiarac?invUa infemalis, xav. From above. xlO. MOXOailAPIt OF JAPANESE OPHIUIIOIUEA, 825 are rather small, and often covered with granules on their inner borders. One of the specimens from Okinawa notably departs from the typical ones. It is only 6 mm. across the disk. The radial shields are almost covered with granules. Moreover, in the arm bases, granules which are finer than the disk granules occur on the ventral abradial parts of the lateral arm plates, on the proximal parts of the same under the arm spines, along the dorsal border of the same, and along the distal border of the dorsal arm plates. OphiavaehneUa megalaspis Clark. Opliiararhiella megalasj^is : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 25, fig. 2. Off Suno Saki, Sagami Sea ; 44-50 fatlioms (Claek). Eastern fSea ; 95-106 fathoms (Clark). Family 4. Ophiochitonidae ^Iatsumoto, 1915. Disk covered with fine, imbricating scales, or rarely with fijio granules ; oral angles always free of granules. Five or six oral papillcT on either side of each jaw ; the outermost one is pointed inwards, projecting above the next papilla, which is tlie largest. Teeth arranged in a single vertical row, either triangular and obtusely pointed, or quadrangular, with very stout, truncated end. Dental papilla) absent. Peristomial plates double or triple. Oral frames with or without well developed lateral wings. Radial shield and genital plate of the same side of a radius articulating with each other by means of two articular condyles and one articular pit. Arms long and slender, widest at some distance from the base, horizontally flexible. Dorsal, lateral and ventral arm plates 326 ART, 2. — H. MATSÜMOTQ : all well developed, with cereous lustre. Two to four, usually three, arm spines, moderately long, erect. One or two large, leaf-like tentacle scales on the abradial border of each pore ; sometimes one to three very small accessory ones may occur on the adradial border. This family includes five genera, whicli may be grouped into two subfamilies as follows. Subfamily 1. Oplnoclntoninœ Maïsumoto, 1915. — Oral frames without well developed lateral wings ; teeth triangular and obtusely pointed ; dorsal surface of vertebrae entire, rhomboidal. Ophiochiton Lyman, 1878.^^ Opliioplax Lyman, 1875. Subfamily 2. Ophionereidince (Ljungman, 1867) mihi, 1915. — Oral frames with well developed lateral wings ; teeth very stout, quadrangular, with truncated end ; dorsal surface of vertebrae notched at the inner end and V-shaped. Ophiodoris Kœhler, 1904. Ophionereis Lütken, 1859. Ophiocrasis Claek, 1911. As shown in the foregoing diagnoses of the subfamilies, there are two types of internal structures in the present family. The internal structures of the Ophiochitonînœ are very similar to those of the Opliiodermatidce. In Ophiochiton fastigatus Lyman, the peristomial plates are triple, one of the secondary plates filhng up the outer open angle formed by the other two. The oral frames are very long, with V-shaped grooves for the ambulacral ring canal, as in Ophiarachna incrasscda and Bathijpecihiura gofoi. The genital 1) In my opinion, Ophiochiton hjmani Sti'der, 1883, does not belong here, and is an Ophiocten, allied to sucli sj^ecies as Ophioden hastatmn Lymax, 1878, and pacificum Lütken & MonTENsEN, 1899. MONOGEAPH OF JAPANESE OPHIUROIDEA. 327 plates and scales are long and slender, being longer and more slender than in the Ophiodermatidœ. The radial shield and genital plate of the same side of a radius articulate with each other by means of two articular condyles and one articular pit. The dorsal surface of the vertebrae is rhomboidal and slightly notched at the inner end but not so strongly as to be V-shaped. In Ophioplax lamellosa Matsumoto, the peristomial plates are also triple, being however relatively shorter and wider than in the preceding species ; the unpaired secondary plate is very small and is slightly notched on the outside. The oral frames are short and much more flared than in the preceding, while the I'oof of the first oral tentacle pore is less projected adradially. In other respects it is almost similar to the preceding. In the Ophîonereidinœ, the internal structures approach those of the next family, Ophiocomidœ, in certain points. In Ophionereis annulata (Le Conte) and 0. reticulata Lütken, the peristomial plates are double, without the unpaired secondary plate, and mucli smaller than in the Ophiochilonince. The oral frames are very long, stout, with well developed lateral wings for the attachment of very voluminous chewing muscles. The teeth are very stout, quadrangular, with widened and trimcated end. The dorsal surface of the vertebra3 is strongly notched at the inner end corresponding to the elongated outer end of the preceding vertebra, so as to be markedly V- shaped. The radial shields, genital plates and scales are fundamen- tally similar to those of Ophiochiton fasfigatus, though the radial shields are much smaller. The internal structures of Ophiocrasis marktannerl Matsumoto are similar to those of Ophionereis, except the peristomial plates, which are triple, with a small unpaired secondary plate. Now we come to a most interesting genus, OphiodoriSy which is fairly intermediate in its internal structures between Ophiochitonince and the Ophionereis- Ophiocrasis-iy^e. In OpJiiodoris 328 ART. 2. — H. MATïUMOTO : pericalles Clark, the peristomial plates are double, without the unpah'ed secondary plate. The oral frames are not very stout and ^have more or less distinct lateral wings, which are less well developed than in Ophionereis and Ophiocrasis. The dorsal surface of the vertebrœ is notched at the inner end so as to be fairly Y- shaped, but not so strongly as in Ophionereis and Ophiocrasis. The radial shields, genital plates and scales are similar to those of Ophionereis and Ophiocrasis. As to the biological significance of the characters of the present family, we can recognise two lines along which advance has been made from the Ophiochitonime to the Ophionereidinœ. In the first place, the Opl doner eidince have better developed chewing apparatus than the Ophiochitonince, such as stouter teeth, with widened and truncated end, smaller peristomial plates, stouter oral plates with well developed lateral wings and more voluminous muscles. In the second place, the arms of the Opliionereidince are more flexible than those of the Ophiochitonince, as may be judged from the notched, V- or Y-shaped dorsal surface of the vertebrae. It may be remarked that the arms of Ophionereis and Ophiocrasis are strongly flexed in alcohol and very freely mobile in life. Key to genera of Ophiochitoninœ, A — Disk entirely free of grannies ; arms markedly keeled above and below Ophiochiton. A A — Disk or at least the interbrachial ventral surfaces covered with fine granules ; arms not keeled, at least below Ophioplax. Ophiochiton fastigatus Lyman. OpMocMton fastigatus : Lyman, Bull. Mus. Comp. Zool., V, 1878, p. JIONOGRAPH or JAPANESE OPHIÜßOIDEA. 329 132, PL YII, figs. 182 & 183; Lyman, Kep. Challenger, V, 1882, p. 176, PI. XXIV, figs. 13-15 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 133. OpMochiton carinatus : Lütken & Mortensen, Meni. Mus. Comp. Zool., XXm, 1899, p. 164, PI. XIV, figs. 1-3. Numerous specimens ; oft' Misaki, Sagami 8ea. Sagami Sea; 110-405 fathoms (Clark). Uraga Channel ; 70- 197 fathoms (Clark). Suruga Gulf; 94-270 fathoms (Clark). Eastern Sea ; 406 fathoms (Clark). Fig. 9L OpMochiton fadigalus. a. From above. x2. h. From below. X 2. c. From above, x 7. d. Yentral view of an oral angle, showing the presence of an accessory oral shield, x 2. e. Side view of three arm joints near tlisk. x2. f. Side view of three arm joints near disk of a medinm- sized specimen, x 3. q. Side view of three arm joints near disk, x 7. c. and g. young. 330 AKT. 2.' — H. MATSUMOTO : Lat. 5° 41' S., long. 134^ 4' E. ; 800 fathoms (Lyman). Gulf of Panama ; 322-546 fathoms (Lütken & Moetensen). The largest of my specimens is 31 mm. in the disk diameter, 400 mm. in the arm length, and 4 mm. in the arm width ; and the smallest one 8 mm. in the disk diameter, 52 mm. in the arm length, and 0.8 mm. in the arm width. This species is very variable, and has been remarked upon by Claek. The adorai shields are sometimes very feeble, but often very w^ell developed, especially in the younger specimens, so as to lie between the oral shields and the first lateral arm plates. Therefore, Kœhlek's subdivision of this genus according as the oral shields are separated or not from the first lateral arm plates, has no meaning at least for the present species. Often, more or less prominent accessory oral shields occur in some of the interradii. This is another fact showing that the present genus is related to Ophiarachna. I consider 0. carinatus Lütken et Moetensen to be a synonym of this species, because specimens of the latter of corresponding size agree almost perfectly with the description and figures of the former. The figures here given, together with those of medium- sized specimens by Lütken & Moetensen will give a fair idea of the variability of this species. Ophioplax lamellosa Matsumoto. Oj^liioplxxx lamellosa : Matsumoto, Proc. Acad. Nat. 8ci. Philadelphia, 1915, p. 88. One specimen ; off" Kôtsu-jima, Sagami Sea. Diameter of disk 4.5 mm. Length of arms 35 mm. Width of arms at base 0.8 mm. Disk flat, thin, pentagonal or rather five-lobed, with indented interbrachial borders, covered with fine, imbricating scales, among MONOGEAPH OF JArANESE OPHIÜROIDEA. 331 whicli the six primaries are somewhat distinct. Central plate rather large, circular, conspicuous. Radiais smaller and less con- spicuous than the central plate, separated from one another and from the central plate. liadial shields triangular, with acute inner Fig. 92. Ophloplax lamellossa. x 16. a. From above, h. From below. c. Side view of three arm joints near disk. angle, twice as long as wide, those of a pair widely separated from each other. Interbrachial ventral surfaces entirely covered with very fine granules. Genital slits very large, reaching to the margin of the disk. Oral shields large, triangular, with strongly curved outer border, less curved lateral sides, obtuse inner angle, and perfectly rounded lateral angles ; longer than wide. Adorai shields also 332 AET. 2. — H. MATSUMOTO : large, triangular, long, tapered within to an acute point, where they do not meet. Oral plates naked. I'ive oral papilla) ; outer- most one long, acutely pointed within ; the second is the largest of all, \ery wide and flat ; the others small, somewhat conical, obtuse. Four or five teeth, very large, with rounded ends, except the uppermost one, which is longer than the others and pointed. Arms long, slender, thin. Lamellar plates and line granules occur on the dorsal and lateral surface of 1:he arm bases. First one or two dorsal arm plates small, partly covered over by the lamellar plates. Those beyond triangular at tlie outset but be- coming quadrangular a little further out, with rounded outer corners and curved lateral borders, wider than long, wider without than within. Lateral arm plates not very prominent. First ventral arm plate small, triangular, with rounded angles, nearly as wide as long. Those beyond pentagonal, with an inwardly directed angle, which is covered by the preceding plate ; outer border curved, lateral borders concave and l^ounded by the tentacle pores. Three arm spines, conical, tapering, obtuse, a little longer than the corresponding arm joint, subequal, but the middle one slightly longer than the other two. One tentacle scale on the abradial side of the pore, very large and oval. Besides, on the adradial side of a few basal pores, there occur one or two rudimentary tentacle scales, more or less covered over by the abradial scale. Colour in alcohol : yellowish gray above and white below ; arms banded with dark gray. It is recorded that the above specimen was taken with a coral net, but the depth is not stated. As the annulation on the arms indicates, this species is not a deep water form, but probably sublittoral. IMONOGKxM'H or JAPANESE OrHIUIlOlDEA, 333 Opliioplax lamellosa is quite near to both 0. Ijungmani Lymak, 1875, and custos (Kœhlee, 1896), as well as to Ophiopeza reduda^^ Kœhlee, 1907. It is distinguished from 0. Ijungmani by the presence of the primary plates, by the coarser disk scales, by the shape of the radial shields, by the disk margin being not so closely granulated, and by the shape of the oral shields ; from 0. custos by the presence of tlie primary plates, by the coarser disk scales, by the shape of the radial shields, by the adorai shields not meeting within, by the shape of the first and second ventral arm plates, by the presence of the lamellar plates at the arm bases, and by the shape of the dorsal arm plates ; and from 0. reducta by the dorsal surface of the disk being free of granules, by the radial shields not being divergent, and by the shape of the ventral arm plates. Kœhler states that, in his specimens of 0. custos the dorsal surface of the disk was also closely covered with fine granules during life, but that they subsequently dropped oif. Whether the same holds true for 0. lamellosa can not be ascertained, as I have not been able to examine living specimens. Key to genera of OpJnonerekUnœ, A — Accessory dorsal arm plates absent ; a row of spines present along the brachial borders of the disk, jnsfc above the arm bases OpJiiodoris. ^^— Accessory dorsal arm plates present. a — Only two large accessory plates to each dorsal arm plate OpJiionereis. 1)' Opldopeza reducta, which Ci.Ar.ft refers to his genus Bathyjjcctinura, appears to Die to be merely an Ophioplax. The presence of only three long, cylindrical arm spines, and of only five papilLo, the annulation on the arms, and the naked oral plates, are all characters of Ophioplax, and not of the genuine Bothilpectinvra. 334 AET, 2. — H. MATSUMOTO : aa — Very small, secondary dorsal accessory xDlates present besides the two large ones OpMocrasis. OpliiodOTis pericalles Clark. Opliiodoris pericalles : Clark, Bull. U. S. Nat. Mus., LXXV, 1911, ]). 177, fig 80. Off Osé Zald, Suruga Gulf; 65 fathoms (Clark). Eastern Sea ; 95-106 fathoms (Clark). Key to Japanese species of OpMoneveis.^^ A — Genital papillas x^i'esent ; ventral arm plates without median notch and eminence. (( — Disk scales coarse ; dorsal arm plates not very wide, slightly wider than long ; three arm spines porrecta. aa — Disk scales very fine ; dorsal arm plates very wide, twice as wide as long or wider ; four arm spines in the basal arm joints eurybracJdplax. A A — Genital papillae absent ; ventral arm plates with a median notch and eminence sinensis. Ojyliioueveis porrecta Lyman. OpMonereis porrecta : Lyman, Proc. Boston Soc. Nat. Hist., VII, 1860, p. 260-) ; Lyman, 111. Cat. Mus. Comp. Zool., I, 1865, p. 147, figs. 14 & 15 ; Lyman, Rep. Challenger, V, 1882, p. 161; Brock, Zeitschr. wiss. Zool., XLVIII, 1888, p. 495; Kœhler, Bull. Sei. Fr. Belg., XXXI, 1898, p. 74; KcEHLER, Exp. Siboga, XLV, Pt. 2, 1905, p. 53. 1) CiiAKK, 1915, lias recorded OpMonereis dubia from •' Japan ; mouth of the Bay of Yeddo " ( = Gulf of Tokyo). If his record is based upon Prof. Morse's specimens, it is probably not 0. dubia but Opldocrasis marldanneri. Some specimens of the last mentioned species in the Zoological Institute, Imperial University of Tokyo, were also labelled " OpMonereis dtibia var." by Rrof. Morse. I have yet to find a specimen of 0. dubia from the Aicinity of the mouth of the Gulf of Tokyo, though OpMocrasis marJdanneri is very common there. 2) This paper was not seen by me. MOXOGEArn OF JAPANESE OPHIUROIDEA. 335 Ophiouereis squamata : Ljungman, Oph. viv., Of. -K. Akad. Forh., 1866, p. 310. Opldonereis variegata : Duncan, Journ. Linn. Soc. Loudo]i, XIV, 1879, p. 462, PI. X, figs. ]5 & 16. Two specimens ; Okinawa. One specimen ; locality unknown. Okinawa, Eiu-kiu. Korea Strait ; 33 fathoms (Duncan). Honolulu. Fig. 93. Opldonereis porrecta. a. From above. xl5. h. From below. xl5. c. Young specimen; from above. x30. d. Side view of three arm joints near disk ; two sets of arm spines removed to show the accessory scales, x 15. 336 ART. 2. — IL MATSUMOTO : The largest one of my specimens is 5 mm. across the disk, with the arms four or five times as long. The apical pair of oral papilla? are infradental, and there is no unpaired papilla. The arms are lightly keeled both above and below, so that they remind us of Ophioc.hiton. There are two or three scale-like sup- plementary plates between the lateral arm plates ; but they are not very conspicuous, being covered over by the arm spines. In smaller specimens, the disk squamation much resembles that of Ophioiona and Ophiozonelld. Ophionereis euvyhvachiplax Clark. OpMonercis euryhracJdplax : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 173, fig. 78. Off Kinkwa San ; 31-41 fathoms (Claek). California. Ophionereis sinensis (Duncan). Ophionereis dubia var. sinensis : Duncan, Journ. Liuu. Soc. London, XIV, 1879, p. 464. Korea Strait ; 33 fathoms (Duncan). This species is very imperfectly described by Duncan as a variety of 0. dubia (Müllee & Teoschel, 1842). His remarks read as follows. " The form from the Korean Sea is well grown, and differs from the type {0. dubia) as follow: — The lower arm-plates have a median notch and eminence ; the spines are subequal, and they are rarely banded with colour. It has a marsupium, and doubtless, as was commonly the case in these Korean species, it is vivip- arous." MONOGBAPH OF JAPANESE OPHIUROIDEA. 337 If these differences between the Korean form and the typical 0. dubia really exist, then it is almost certain that the former is not referable to the same species. Therefore, I am inclined provisionally to look upon the former as a distinct species. Ophiocrasis Claek. 1911. Aside from the presence of the secondary supplementary dorsal arm plates, this genus seems to me to be distinguished from Ophionereis principally by negative characters and the diflerent degree of development of certain common structures. The disk scales are even and exceedingly fine ; no trace of the marginal row of special disk scales ; no genital papillse ; arms much nar- rower than in Ophionereis. Schizogonic reproduction may not be a generic character. Key to species of Ophiocrasis» A — Kadial shields not very small ; ventral arm plates longer than wide throughout ; smaller individuals sis-armed, reproducing by schizo- gony diciydisca. AA — Radial shields exceedingly small and insignificant ; ventral arm plates wider than long, at least in the basal parts of the arms ; five-armed, not reproducing by schizogony marhtanneri . Ophiocrasis dictydisca Claek. Ophiocrasis dictydisca : Claek, Bull. U. S. Nat. Mus., LXXV, 1911, p. 175, fig. 79. Off Suno Saki, Sagami Sea ; 52-73 fathoms (Claek). Korea Strait ; 59 fathoms (Claek). 338 ABT. 2. — H. MATSUMOTO : Ophiocrasis marktanneri Matsumoto. OpUonereis porreda : Marktanner-Tüeneretschee, Ann. K. K. Naturli. Hofmus., II, 1887, p. 302, PL Xn, fig. 18. (Non Lyman, 1860.) ' OpUonereis porreda Marktanner ': Kœhler, Bull. Sei. Fr. Belg., XXXI, 1898, p. 76. Ophiocrasis marManneri : Matsumoto, Proc. Acad. Nat. Sei. Philadel- phia, 1915, p. 90. Three specimens ; Enoshima, Sagami. Numerous specimens ; Arai Beach, Misaki Marine Biological Station. Diameter of disk 9 mm. Length of arms 68 mm. Width of arms at base 1.2 mm.; at the widest part 1.5 mm. Disk circular, slightly convex, rather soft, covered with very fine, imbricating scales, which are rather obscured, so that the disk appears as if covered by a thick skin. Radial shields very small, short and exceedingly narrow, tapered within, those of a pair widely separated from each other, also rather obscure. Interbrachial ventral surfaces cover- ed with scales similar to those of the dorsal side, but even more obscure. Genital slits large, nearly reaching to the border of the disk. No genital papillae. Fig. 94. Ophiocrasis marktanneri. x7. a. From above. h. From below, c. Side view of three arm joints near disk. MONOGEArH OF JAPANESE OPHIUROIDEA. 339 • Oral shields rhomboidal, with obtuse inner angle, rounded lateral angles, and widely rounded outer angle ; nearly as wide as long, except the madreporic shield, which is decidedly longer than wide. Adorai shields small, acutely tapered within, where they nearly or hardly meet. Four or five oral papilloo on either side, unequal, short, rounded, but the outermost one, which is closely associated with the second oral tentacle pore, has pointed inner end. Four teeth, short, very wide and stout. Arms long and very slender, narrowed at the base, widest at one-fourth to one-third the arm length from the base. Dorsal arm plates mostly triangular, with obtuse outwardly directed apex, rather small, wider than long, successive plates slightly in contact ; quadrangular in the more distal parts. On either side of each dorsal arm plate, occurs a large supplementary plate, which is nearly semicircular, about one-half as large as the dorsal arm plate, and bounded along the distal border by one or two very insignificant secondary pieces, which however are present only for a comparatively short extent. Two or three basal dorsal arm plates and their supplementary plates are smaller than those beyond. The supplementary plates become smaller outwards as the dorsal arm plates become quadrangular, and finally disappear. Lateral arm plates not very prominent, not meeting above or below. First ventral arm plate very small, rather pentagonal, longer than wide. Those beyond quadrangular, with rounded outer lateral angles, truncated inner lateral angles, and slightly notched outer border, about as long as wide, but longer than wide outwards. Three arm spines, short, stout, flattened, slightly tapered, blunt. One large and oval tentacle scale. Colour in alcohol : grayish yellow ; disk reticulated, and arms banded, with dark purplish brown. 340 AET. 2. — H. MATSUMOTO : • The arm length varies from six to eight times the diameter of the disk. In the smaller specimens, the arm spines are less flattened ; and in those of less than about 4 mm. across the disk, the secondary supplementary pieces on the arms are almost invisible. This species differs from the genotype, 0. didydisca Claek, in the shape of the dorsal arm plates, in the less well developed secondary supplementary pieces, in the smaller and less distinct radial shields, and in the smaller arm spines of the basal joints. Further, schizogonic reproduction has not been observed in this species, though I have examined many small specimens. On the other hand, 0. marktanneri, as well as the genotype, resembles Ophionereis dubia (Müllek & Teoschel, 1842) in lacking the genital papillœ, but differs from it chiefly in the presence of secondary supplementary dorsal arm plates, and in the much narrower arms. 0. marktanneri is by no means near to Ophionereis porrecta Lyiman. I could mention some more differences than those enumerated by Kœhlee between these two species, but it will not be necessary to do so here. This beautiful species is one of the most common ophiurans about Misaki, living under stones and rocks. Family 5. Ophiocomidae Ljungman, 1867. Disk covered with fine granules or by a naked skin, some- times bearing scattered spines. Radial shields very stout, but externally invisible, those of a pair widely separated from each other. Four to six oral papilla3 on either side, the outermost one being pointed inwards above the next one. Dental papillae well developed, forming a vertical clump at the apex of each jaw. Teeth quadrangular and very stout. Peristomial plates double. MONOGRAPH OF JAPANESE OPHIUEOIDEA. 341 Oral frames very stout, with extremely well developed lateral wings. The radial shield and genital plate of the same side of a radius articulate with each other ])y means of two articular condyles aud one pit. Arms moderately long, stout, widest at some distance outside the base. Arm plates all well developed. Arm spines long, stout, perpendicular to the arm axis ; the uppermost spines of every second or tliird lateral arm plates often very large and clavate. One or two tentacle scales to each pore. This family includes five genera, four of which form a sub- family, while the fi.fth forms another. Subfamily 1. Ophiocomince Matsumoto, 1915 : — Eadial shields very stout, boot- shaped, those of a pair widely separated from each other ; three to five arm spines, the uppermost spines of every second or third lateral arm plates being usually very large and club-shaped ; tentacle scales short and leaf-like. Ophiopteris Smith, 1877. Ophiocoma Agassiz, 1835. Ophiomastix Müller & Troschel, 1842. Ophiarthrum Peters, 1851. Subfamily 2. OphiopslUnœ Matsumoto, 1915: — Eadial shields slender, bar-like, each pair being rather close together at the outer ends corresponding to the dorsal approximation of the outer ends of the genital plates ; numerous arm spines, none of which is club-shaped ; two tentacle scales, of which the adradial one is very long and lanceolate, while the abradial one is very short and acute. Ophiopsila Forbes, 1842. The present family is very uniform in its internal structures. I have dissected all the Japanese representatives of this family, viz., Ophiocoma hrevipes Peters, 0. scolopcndrina (Lamarck), Ophio- 342 AKT. 2. — H. MATSUMOTO : mastix mixta Lütken, 0. ammlosa (Lamaeck), 0. lütkeni Pfeffer and Ophiarthrum elegans Petees. In all of them, the radial shields are very stout and boot- shaped, and each consists of a radial bar and a transverse vring projecting laterally from the outer part of the former. Each shield has two articular condyles and one pit, fitting to the two condyles and one pit of the corresponding genital plate. The latter is bar-like, slightly curved laterally, articulating with the genital scale at some distance inwards from the outer end. The genital scales arc also b9,r-like, more or less flattened, but not so wide and leaf -like as those of the Gnathophiurida. The peristomial plates are rather small and double, the two being firmly united together. The oral frames are exceedingly stout, with extremely well developed lateral wings for the attachment of very voluminous masticatory muscles. The oral and dental plates taken together are "IT- shaped in dorsal view, the former completely overlapping the latter. They are quite different from those of the Gnathophiurida. The teeth are quad- rangular and very stout. In Ophiopsila riisei Lütken, 1859, the internal structures are almost similar to those of the preceding, save the radial shields, which are slender and bar-hke, without transverse wings, and rather closely set in pairs. Ophiopsila is by no means near to the Amphiuridce in internal structures, so that I have no hesita- tion to refer it to the present family, forming however a sub- family for it alone. Key to Japanese genera of Ophiocomidœ, A — Disk covered with granules or bearing spines. a — Disk covered only with granules, entirely free of spines . . OpJdocoma. MONOGRAPH OF JAPANESE OPHIUROIDEA. 343 aa — Disk bearing spines, or both granules and spines .... Opidomastix. AA — Disk covered by a naked skin, entirely free of granules and spines Ophiarthrum. Key to Japanese species and varieties of Ophioconia, A — Interbracliial ventral surfaces entirely covered with granules ; oral shields usually oval; dorsal arm plates with clearly cut outer border ; arm spines shorter dorsally than ventrally hrevipes. AA — Granulations of the interbrachial ventral surfaces not extending to the abradial border of the genital shts, the inner parts being mostly naked ; oral shields quadrangular ; outer border of the dorsal arm plates not clearly cut ; arm spines longer dorsally than ventrally . . scolopendrina. a — Disk variegated; arm spines annulated typical scolopendrina. aa — Disk as weU as arm spines uniformly black. h — Arm spines not very short and stout ; two tentacle scales to each pore var. erinaceus. hb — Arm spines exceedingly short and stout ; a single tentacle scale to each pore, except in a few basal arm joints var. schœnUinii. Ophiocofiia brevipes Petees. Ophiocoma hrevipes : Peters, Arch. Naturg., 1852, j)- 85^^; Lyman, Rep. Challenger, V, 1882, p. 172 ; Marktanner-Turneretschee, Ann. K. K. Natm-h. Hofmus. Wien, II, 1887, p. 303 ; Bell, Proc. Zool. Soc. London, 1888, p. 388; Döderlein, Zool. Jahrb. Sys., Ill, 1888, p. 831; Loriol, Mém. Soc. Phys. d'Hist. Nat. Genève, XXXH, 1894, p. 25, PL XXIH, fig. 4 ; Döderlein, Semon - Zool. Forschungsr., V, 1896, p. 289 ; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 61 ; Kœhler, Bull. Sei. Fr. Belg., XLI, 1907, 1) This paper was not seen by me. 344 A1£'J\ 2, — II. MATSUMOTO : p. 325 ; Clauk, Bnll. Mus. Comp. Zool., LI, 1908, p. 296 ; (Jlaiîk, Bull. U. S. Niit. Mus., LXXV, 1911, p. 250. Throe specimens ; Natsui, Iliuga. One specimen ; Okinawa. One specimen ; Yayeyama, Riu-kiu. Three specimens ; Koshun, Formosa. Tln^ee specimens ; locality unknown. Indo-Paciflc. Natsui appears to be the northern limit of this species for the present. The colour in alcohol is very variable. In some specimens, the disk is dark grayish brown and sparsely spotted with black. In others, tlie disk is light bluish green and variegated with white. The oral shields are nearly oval in some specimens, but more or loss oblong quadrangular in others. 'l'Ile disk granules are very lino and close-set. The granules occur also on tlie abradial end Fig. 9S. Oplàocoma hrevipeN. xd. u. From above. I>. From below. c. Side view of three urm joints near disk. of the adorai shields and of the first lateral arm plates. The oral })apilloo are comparatively numerous and number five or six on either side. The dorsal brachial margins of the disk are concave. The dorsal arm plates are wide and have clearly cut outer borders. The arm spines are shorter dorsally than ventrally. A few of the lower dental papillaa are very rudimentary and granule- like, reminding us of the granules of the oral angles of tho MONOGR-Vril OF JAPANJ'ISE orillUriOIDEA. del Ophloclermatidce. Many of these characters indicate tliat the present species is a most primitive form of Ophiocoma. Ophioconia scoloxtendrlna (La:siarck). Ophiocoma scolopendrina var. crinaceiis (MürJjER & Teoschel). OpJdocoma scolopendrina var. schœnleinii (MülIjEU & Trosohel). OpMura scolopendrina : Lathaeok, Hist. Nat. Anira. sans Vert., II, 181G, p. 54. OpJdocoma scolopendrina: Agassiz, Mem. Soc. Nat. Noucliatol, I, 1835, p. 192'^; Lyman, Rep. Challenger, V, 1882, p. 170 ; Studee, Abh. K. Preuss. Akad. Wiss. Berlin, 1882, p. 20; Makktanner-Turneretscher, Ann. K. K. Naturh. Hof mus. Wien, II, 1887, p. 302 ; Bell, Proc. Zool. Soc. London, 1888, p. 388; ]îrock, Zeitschr. wiss. Zool., XLVII, p. 495; Döderlein, Zool. Jahrb. Sys., HI, 1888, p. 841 ; Loriol, liev. Suisse Zool., I, 1893, p. 407; Loriol, Mém. Soc. Pliys. d'llist. Nat. Genève, XXXH, 1894, p. 23; Doderlein, Semon-ZooI. Porschungsr., V, 189G, p. 288; Piaffer, Abh. Senckenberg. Naturf. Gesell., XXV, 1900, p. 83 ; KaiHLER, Exp. Siboga, XLV, Pt. 2, 1905, p. GO; Kœhler, Bull. Sei. Fr. Bclg., XLI, 1907, p. 32G; Clark, Bull. Mus. Comp. Zool., LI, 1908, p. 297. OpJdocoma crhiaccus : MÜLLER & Trosciiel, Sys. Ast., 1842, p. 98 ; Lyman, loc. cit.; Studee, loc. cit.; Marktanner-Türneretscher, loc. cit.; Brock, loc. cit.; Loriol, loc. cit., 1893, p. 419; Loriol, loc. cit., 1894, p. 21; Clark, loc cit., p. 296. OpJdocoma erinaced^^: Doderlein, loc. cit., 1888, p. 289; Clark, ]îull. U. S. Nat. Mus., LXXV, p. 257. OpJdocoma scolopendrina var. erinaceus : Dodellein, loc. cit., 189G, p. 289 ; Kœhler, loc. cit., 1905 ; K(EHLER, loc. cit., 1907. OpJdocoma scJiœnleinii : Müller & Trosciiel, loc. cit., p. 99 ; Lyman, 1) This palmer was not seen by mo. 2) I think Müller & Teoschel used the spocific name as an appositivo noun and not as an adjective, ei'lnaceus signifying a hedgehog; so that, there is no reason to change it to "erinacea." 346 ART. 2, — H. MATSUMOTO 111. Cat. Mus. Comp. Zool., I, 1865, p. 70 ; Lyman, loc. cit., 1882, p. 171 ; Claek, loc. cit., 1908, p. 296. One specimen ; Pinnacle Is. Four specimens ; Bonin Is. Numerous specimens ; Botel Tobago. Numerous specimens ; Koshun, Formosa. Four specimens ; Yayeyama, Riu-kiu. Six specimens ; Osnima, Osumi. Numerous specimens (belonging to the Seventh High School, Kagoshima); Kagoshima Gulf. Indo-Pacific. Kagoshima Gulf seems to be the northern limit of this species. There is no doubt that 0. scolopendrina, erinaceus and schœnleinii are conspecific, as rightly noted by Maekt année - TüRNEKETSCHER. I havo closcly examined many intermediate specimens, and am fuUy convinced that there is no line of demarcation. The continuity of the scolapendrina- and erinaceus- type has already been remarked upon by IjUdwig, Maektanner- Fig. 96. Opldocoma scolojxndr'ma Tar. erinaceus. a. From above. X 3. h. From below, x 3. c. Side view of three arm joints near disk. x3. d. Young specimen from above. xl4. MONOGRAPH OF JAPANESE OPHIUROIDEA. 347 TuRNEEETSCHER, DÖDERLEIN and Kœhler. Typical specimens of the schœnleinii-type have very short and exceedingly stout arm spines and only one tentacle scale to each pore, except in a few basal arm joints witliin the disk. But there are also many specimens of the ermaceus-ty-pe with only one tentacle scale, instead of two, to each pore in most of the arm joints. Now, there are two classes of specimens which are intermediate between the erinaceitS' and schœnleinii-iypQ. In the first, the arm spines are very short and stout, and there is mostly only one tentacle scale to each pore, though often two are present in some of the free arm joints. In the second, the arm spines are not very short and stout, and there is only one tentacle scale, except in a few basal arm joints within the disk. There are also some specimens which are intermediate between the scolopendrina- and schœnleinii-ijpe, in which the arm spines are very short and stout, and not uniformly blackish, as in the schœnleinii- and erinaceus'typc:, but annulated or spotted with white as in the scolopendrina-ij^Q, while of the tentacle scale there is mostly only one, though two may be present in some of the arm joints. The numbers of the specimens of each type and of the intermediate ones, together with the localities, are shown in the following table : -^.„^Localities : Types :^ ~----.-_^^ Bonin Is. Botel Tobago. Koshmi. Yayeyama. Okinawa. scolopendrina 2 8 7 1 2 intermediate 1 3 6 0 0 erinaceus 1 9 7 1 1 intermediate 0 4 2 0 1 scJiœnleinii 0 7 1 0 2 intermediate 0 1 2 2 0 scolopendrina — — — — — 348 ART. 2. — H. MATSUMOTO : Key to Japanese species of Ophiomastix, A — Disk closely covered with granules and bearing a number of very short spines ; dorsal arm plates fan-shaped ; colom' dark red in life and whitish in alcohol mixta. A A — Disk sparsely covered with spines, or with spines and granules, neither being very close-set. a — Disk sparsely covered with both spines and granules ; dorsal arm plates hexagonal ; colour black IVdkeni. aa — Disk sparsely covered Avith spines only ; dorsal arm plates rhom- boidal ; colour grayish brown, disk spines and arm sxjines finely annulated with white cinnulosa. Oxyhioniastix fnixta Lütken. Ophiomastix mixta : Lütken, Add. Hist. Oph., Pt. HI, 1869, p. 44 ; Lyman, 111. Cat. Mus. Comp. Zool., YI, 1870, p. 15; Lyman, Rep. Challen- ger, V, 1882. p. 175; Brock, Zeitschr. wiss. Zool., XLVII, 1888, p. 497; LoRiOL, Rev. Suisse Zool, I, 1893, p. 414 ; Kœhler, Exp. Siboga, XLV, Pt. 2, 1905, p. 68, PI. YI, fig. 15, PI. XV, fig. 1 ; Clark, Bull. U. S. Nat. Mus., LXXV, 1911, p. 256, fig. 126. Fig. 97. Ophiomastix mixta. x3. a. From above, b. From below, c. Side view of three arm joints near disk. MONOGEAPH OF JAPANESE OPHIUROIDEA. 349 One specimen ; Enoshima. Numerous specimens ; Misaki. Indo-Pacific. Misaki may be the northern hmit of this species. My specimens have the disk spines not so profusely developed as in the East Indian type, but fevv^, short and inconspicuous ; so that they are rather liable to be mistaken for an Ophiocoma. The colour is yellowish white in alcohol, but very dark red in life. Ophiomastix liltkeni Pfeffer. Ojoh'omastix liltkeni : Pfeffer, Abb. Senckenberg. Naturf. Gesell., 1900, XXV, p. 83. Four speci- mens ; Okinawa. Philippines (Pfeffer). The largest one of my speci- mens is 22 mm. in diameter, 250 mm. in the arm length, and 4 mm. in the arm width at the base. These specimens agree exactly with Pfeffee's des- Fig. 98. Ophiomastix lütkeni. x3. a. From above, b. From below, c. Side view of three arm joints near disk. 350 AKT. 2. — H. MATSUMOTO cription, if we make what I consider to be necessary corrections in his expressions : the three rows of teeth mentioned by him are in reahty dental papillae ; and above them, deep within the mouth, there is a row of about four teeth, which are hardly visible without dissec- tion. In larger specimens, the uppermost stout arm spine is often bifurcated at the free end. The colour is black as in Pfeffer' s type. OpJiioniastix annulosa (Lamarck). Ophiura annvlosa : Lamaeck, Hist. Nat. Anim. sans Vert., IT, 1816, p. 543. Opiiiomasiix an- nulosa : Müller & Troschel, Sys. Ast., 1842, p. 107; LÜT- KEN, Add. Hist. Opli., ni, 1869, p. 44; Lyjian, in. Cat. Mus. Comp. Zool., VI , 1869, p. 15 ; Lyman, Rep. Challenger, V, 1882, p. 175; Stüder, Abh. K. Preuss. Akad. Wiss. Berlin, 1882, p. 21 ; Brock, Zeit- schr. wiss. Zool., XLVII, 1888,p.497; DÖDERLEIN, Zool. Jahrb. Sys., HI, 1888, p. 832 ; LoRiOL, Eev. Suisse Zool., I, 1893, p. 413; Kceh- Fig. 99. Ophiomastix annulosa. x3. a. From above, h. From below, c. Side view of four arm joints near disk. LER, Mem. bOC. ZiOOl. MONOGEAPH OF JAPANESE OPHIüEOIDEA. 351 Fr., Vm, 1895, p. 403 ; Döderlein, Semon— Zool. Forschnngsr., V, 1897, p. 289, PI. XVI, fig. Ill; Kœhlee, BuU. Sei. Fr. Belg., XXXI, 1898, p. 84; Pfeffee, Abb. Senckenberg. Natnrf. Gesell, XXV, 1900, -p. 85 ; Kœhlee, Exp. Siboga, XLV, Pt. 2, 1905, p. 65; Kœhlee, BnU. Sei. Fr. Belg., XLI, 1907, p. 329; Claek, BuU. Mus. Comp. Zool., LI, 1908, p. 297. Four specimens ; Pinnacle Is., Riu-kiu. One specimen ; Ko- shnn, Formosa. Fom- specimens ; Botel Tobago, Formosa. Indo-Pacific. Pinnacle Is. may be the northern limit of this species, Ophiarth^mm elegans Petees. Opidartlirmn degans : Petees, Monatsb. K. Akad. Wiss. Berlin, 1851, p. 463'>; Petees, Arch. Naturg., XVm, 2, 1852, p. 82^^; Lyman, Hep. Challenger, V, 1882, p. 174 ; Stüdee, Abh. K. Akad. Wiss. Berhn, 1882, p. 21 ; Beock, Zeitschr. wiss. Zool., XLVn, 1888, p. 497; DÖDEELEIN, Zool. Jahrb. Sys., Ill, 1888, p. 831 ; Dödeelein, Semon — Zool. Forschungsr., V, 1896, p. 289; Kœhlee, Buh. Sei. Fr. Belg., XXXI, 1898, p. 108, PI. rn, figs. 25 & 26 ; Ludwig, Abh. Sencken- berg. Naturf. Gesell., XXI, 1899, p. 547; Pfefeee, ibid., XXV, 1900, p. 83; Kœhlee, Ech. Indian Mus., Shallow-wat. Oph., 1890, PI. XIX, figs. 36 & 37 ; Kœhlee, Exp. Siboga, Fig. lOO. Ophiarthrum elegans. x 3. a. From above. ^-^'^' "''• ^} 19ÜO, p. 73; &. From below, c. Side view of three arm joints near Kœhlee, Bull. Sei. Fr. Belg., '''^^'• XLI, 1907, p. 329; Claek, Bull. Mus. Comp. Zool., LI, p. 297. 1) These papers were not seen by me. 352 ART. 2. — H. MATSUMOTO: One specimen ; Okinawa, Riu-kin. Indo-Pacific. Okinawa may be the northern hmit of this species. The Ophiuran Fauna of Japan. H. L. Claek distinguishes four geographical ophim^an faunae in the North Pacific, namely Honshu, Bering, American and Oceanic. I will discuss the ophiuran fauna of . Japan from a somewhat different standpoint from his. All the known Japanese ophiurans are referable to eighty - eight genera, which may be divided from the distributional stand- point into five groups as follows. I. Cosmopolitan genera, which are widely distributed in both the Indo-Pacific and Atlantic, as well as in both the Arctic and Antarctic or either one of the latter. The following genera belong here. Ophiacantha, Amphiactis, Amphiodia, AmphipJiolis, Amphiura, Homalophiura, Ophiiira. 7 genera : ^%. II. Arctic genera, which are essentially circumpolar in dis- tribution or limited to the Bering Sea and very northern Pacific. The following genera belong here. Gorgonocephalus, Ophiolehes, Ophiosemnotes, Ophiopho- lis, Stegophiura, Ophiopenia. 6 genera : 1%. HI. Intertropical genera, which are widely distributed in both the Indo-Pacific and Atlantic, but not in the Arctic or Antarctic. The following genera belong here. Ophioleptoplax, Ophiomyxa, Ophiodera, Ophiohyrsa, MONOGEAPH OF JAPANESE OPHIUEOIDEA. 353 Asteromjx, Asteroschema, Asteroporpa, Ophiomyces, Ophiologimus, Ophiopora, Ophiolimna, Ophiomiirella, OpUacantheUa, Ophiothamnus, Ophiurotliamnus, Ophio- plinthaca, Ophiomitra, Ophiocamax, Amphilepis, Ophi- actis, Amphioplus, Amphichilus, Ophiophragmus, Opino- centrus, Ophiothrix, AmpkiopJmira , OjMomusium, Ophiozonella, Ophiolepis, Ophioleuce, Ophiermis, Ophiu- roconis, Bathypedimira, Pectinura, Ophioplax, Ophio- nereis^ Ophiocoma. 37 genera : i2%. IV. Indo- Pacific genera, which are distributed in the Indo- Pacific, with the centre of distribution chiefly in the vicinity of Malaysia. The following genera belong here. Astroceras, Trichaster, Euryale, Astrocharis, Astro- dendrum, Astrocladus, Astrohoa, Astrothamnus, Astro- toma, Astroclon, Ophiamhix, OpJiiojMmlmus, Ophio- mœris, Amphiacantha, Ophiomasa, Ophiothela, Astro- phiura, Aspidophiura, OpUoplocus, Ophiotroclius, Ophiurodon, Opliiarachna, Ophiarachnella, Ophiochiton, Ophiodoris, Ophiomastix, Ophiarthrum. 27 genera: 30X. V. Honshit genera, which are known only from the vicinity of Honshu, Shikoku and Kiushii. The following genera belong here. Ophiostyracium, Ophiosyzygus, OpUostiha, Ophiohyalus, Ophiohymen, Ophiophrixus, Ophiosmilax, Astrothorax, Ophiophrura, Ophientrema, Ophiocrasis. 11 genera: 13%. With the exception of OpMophthalmus, which is both tropical and Arctic, the Indo-Pacific and Intertropical genera are represen- 354 AET. 2. — H. MATSÜMOTO tatives of tropical fauna, and form 72X of the total, in contrast to the Arctic elements, which amount to only 7%. Moreover, almost all of the Honshu genera have tropical, but not Arctic, affinities. Two hundred and thirty-two species of ophiurans are known from Japanese waters, which may be grouped as follows. I. Circumpolar species, which occur in the North Atlantic, Arc- tic and North Pacific. The species belonging here and their known southern limits in the Nortli Pacific are as follows. Species. Southern limit. Ophiacantha hidentata : oft' southern Korea, Sea of Japan. Ophiopliolis aculeata, typical : off" southern Korea, Sea of Japan. Ditto, including var. japonica : Stegophiura nodosa : Ophiiira sarsii : off Kii, Kumano Sea. off Saghalin, Okhotsk Sea. Eastern Sea ; Uraga Channel (on the Pacific side). 4 species : 2%. IL Bering-American species, which range from the Bering Sea to Japan and to the western coast of U. S. A. The species belonging here and their known southern limits in Japan are as follows. Species. Gorgonocephalus caryi : Ophiophtlialmus normani : Amphiodia euryaspis ; Ophiura cnjptolepis : Ophiura leptoctenia : Southern limit. Eastern Sea ; Suruga Gulf (on the Pacific side). Eastern Sea ; off" Kii, Kumano Sea (on the Pacific side), off" Korea, Sea of Japan, off" Omai Zaki, Yenshti Sea. off" southern Korea, Sea of Japan ; off" Omai Zaki, Yenshù Sea (on the Pacific side). MONOGKArif OF JAPANESE OPHIUßOIDEA. 355 5 species : 2%. III. Bering species, wliicli range from the Bering Sea to Japan, but not to the coast of U. S. A. The species belonging here and their known southern h'mits are as follows. Species. Ophioplithalmus cataleimmoidus : Ophiacantlia adiaphora : Ophiacantha rhachophora : Ophiacanthella acontophora : Ophiosemnotes tylota : Amphiodia cratewdometa : Ophiura quadrispina : Ophiura maculata : Southern limit, off Mikawa, Yenshii Sea. Sagami Sea. Eastern Sea ; off Kii, Kumano Sea (on the Pacific side). off Kurile Is. off Tsugaru, Sea of Japan, off Korea, Sea of Japan, off Shiribeshi, Sea of Japan, off Abashiri, Kitami, Okhotsk Sea. off Sado, Sea of Japan. Ophiopenia dlsacantha : 9 species : 4:%. IV. Intel' tropical species, which are widely distributed in both the Indo-Pacific and Atlantic, but not in the Arctic or Antarctic. The intertropical species usually occur in both the East and West Indies. The species to be referred here and their known northern limits in the North Pacific are as follows. Species. Asteronyx loveni^\' Ophiolimna hairdi : Ophiacantha pentagona^\' Ophiacantha rosea : Ophiactis savignyi^\' Northern limit. Bering Sea. Bering Sea. Sagami Sea. Sagami Sea. Misaki districts. 1) These species have been recorded also from East Indian waters. 356 ART. 2. — H. MATSUMOTO Eastern Sea. off Boshû. off Kinkwasan, Rilmzen. Bering Sea. Uraga Channel. British Columbia. Uraga Channel. off Kii, Knmano Sea. Ampkiophmra convexa : Amphiophiura sculptilis^h Opliiura flag ell ata^^: Ophiura irrorata}^: Ophiomusium scalar ^^: Ophiomiisimn lymam^\- Ophiomusium cancellatum : Ophiernus adspersus^^: 13 species : QX. V. Inch-Pacific American species, which range from Malaysian waters to the western coast of southern North America or Central America, through Japanese waters. The species belonging here and their known northern limits are as follows. Species. Northern limit. Amphiura koreœ : off Echizen, Sea of Japan ; Sa- gami Sea (on the Pacific side). Ophiochiton fastigatus : Uraga Channel. 2 species : 1%. VI. Indo-Pacific species, which range from the Indian Ocean or Malaysian waters to Japan. The species to be referred here and their known northern limits are as foUws. Species. Northern limit. Ophiomyxa aiistralis : Sagami Sea. Astroceras pergamena : Uraga Channel. Trichaster palmiferas : Colnett Strait. Trichaster elegans : Tanabé Bay, Kii. Euryale aspera : Okinawa, Riu-Kiu. 1) These species have been recorded also from East Indian waters. MONOGRAPH OF JAPANESE OPHIUKOIDEA. 357 Astrodendrum sagaminum Astrocladus coniferus : Astroboa nigra : Ophiambix acideatus : Ophiacantha dallasii : Ophiacantha levispina : Ophiacantha inutilis : Ophiomœris obstricta : Ophiactis aßnis : Ophiactis macrolepidota : Ophiactis modesta : Amphiacantha formosa : Amphiura bellis : Amphiura iris : Amphiura liitkeni : Ophiocentrus verticillatus . Ophiothrix koreana : Ophiothrix nereidiria : Ophiothrix obtusa : Ophiothrix hirsuta : Ophiothrix punctolimbata . Ophiothrix longipeda : Ophiothela dance : Aspidophiura forhesi : Ophiura kinbergi : off Shiribeshi, Sea of Japan ; Sagami Sea (on the Pacific side). Wladiostok ; Sagami Sea (on the Pacific side). Hirado Strait. Colnett Strait, off Korea, Sea of Japan, off Shiribeshi, Sea of Japan ; Yen- shti Sea (on the Pacific side). Uraga Channel. Eastern Sea. Korean seas. Uraga Channel. Misaki districts. Strait of Formosa. Sagami Sea. Sagami Sea. Korean seas. Misaki districts. Hakodate Bay. Misaki districts. Okinawa, Piiu-Kiu. Pinnacle Is., Riu-Kiu. precise locality in Japan not known. Kominato, Boshû. Misaki districts. Korea Strait. off Echigo, Sea of Japan ; Uraga Channel (on the Pacific side). 358 ART. 2. — H. MATSUMOTO : Opliiura imbecülis : Ophiomusium simplex : Ophiomusium lunare : Ophiomusium lütkeni : Ophiomusium laqueatum : Ophiozonella projecta : Ophiolepis cincta : Ophiolepis annulosa : Ophioplocus imhricatus : Ophiurodon grandisquama : Ophiarachna incrassata : Ophiaraclmella gorgonia : Opkiarachnella infenialis : Ophionereis porrecta : Ophiocoma brevipes : Ophiocoma scolopendrina : Ophiomastix mixta : Ophiomastix lütkeni : Ophiomastix annulosa : Ophiarthrum elegans : in Japan not Sagami Sea. Eastern Sea. Eastern Sea. Eastern Sea. Eastern Sea. Sagami Sea. precise locality known. Okinawa, Riu-Kiu. Okinawa, Riu-Kiu. Sagaroi Sea. Okinawa, Eiu-Kiu ; Benin Is. Misaki districts. Eno-ura, Suruga Gulf. Korea Strait. Natsui, Hiuga. Kagoshima Gulf. Misaki districts. Okinawa, Riu-Kiu. Pinnacle Is., Riu-Kiu. Okinawa, Riu-Kiu. 50 species : 21%'. VII. Honshu- American species, which occur on both the western and eastern coasts of the North Pacific. The following species belong here. Amphioplus rhadinohrachius. Amphioplus hexacanthus. Amphioplus macraspis. Amphipholis pugetana. Amphiura carchara. MONOGRAPH OF JAPANESE OPHIUROIDEA. 359» Amjr)hmra a cry st at a. Amph iopli Iura inonder osa. Ophiomiis'mm JoUiense. Opliionereis euryhrachiplax. 9 species : ^%. VIII. Honshu species, which are known only from the vicinity of Hon- shu, Shikoku and Kiushû. The following species belong here. Ophiosyzygus disacantlius. Ophioleptoplax megapora. Ophiostyraciwn trachyacanthum. Ophiostiha hideklL Ophioliyalus gotoL Ophlodera anisacantha. Ophiohymen gymnodiscus. Ophiohyrsa acaii thmobrachia . Opliiohyrsa synaptacantha. Ophiophrixus acanthlmis. Ophiosmilax mirabilis. Eur y ale anopla. Aster oschema tub if er urn. Asteroschema glaucum. Asteroschema hern igymnum . Asteroschema caudatum. Asteroschema japonicum. AsteroscJiema ahyssicola. Asteroschema glutinosum. Astrocharis ijimai. Asteroporpa hadracantha. Gorgonocephalus iuherosus. Gorgonocephalus dolichodactylus. 360 AET. 2. — H. MATSUMOTO : Astrocladus annulatus. Astroboa arctos. Astroboa globifera. Astrothamnus echinaceus. Astrothorax misakiensis. Astrotoma sobrlna. Astroclon suensoni. Ophiomyces spathifer. Opliiologimus hexactis. Ophioplwura liodisca. Ophiopora megatrema. Opliiolimna lambda. OphioUmna diastata. OpMomitrella stellifera. Ophiomitrella polyacantlia. Oph ioph th almus leucorh abdotus . Oplilophthalmus codonomorplia. Ophiophtlialmus hylacantha. Ophiophth almus microhylax. Ophientrema eupliylacteum. Ophientrema scolopendricum. Ophlacantha œnlgmatlca. Ophiacantha omoplata. Ophiacantha ancli'ilabra. Ophlacantha lophobrachia. Ophiacantha acanthlnotata. Ophlacantha prlonota. Ophlacantha dlploa. Ophiacantha blsquamata. Ophlothamnus habrotatus. MONOGRAPH OF JAPANESE OPHIUKOIDEA. 361 Ophiothammis venustus. OpJi lurothamnus dlcyclus. Ophiopllnthaca cardlomorplia. Opliiomitra bi/thiaspis. Ophiomitra Uthosora. Ophiocamax polyploca. Ophiolebes brachygnatha. Ophiolebes asaphes. Ophiolebes tuberosa. Ophiosemnotes œcUdisca. Ophiomœi'is projecta. Amphilepis tenuis. Amphiactls umbonata. Ophiactis bracliygenys. Ophiactis jjteropoma. Ophiactis dyscrita. Ophiactis gymnochora. Ophiopholis mirabilis. Ophioph olis br achy actis. Amphioplus megapomus. Amphioplus ancistrotus. Amphioplus cernims. Amphioplus glaucus. Amphichilus trichoides. Amphiacantha acanthina. Amphiacantha dividiia. Amphiodia psilochora. Ophiophragmus japonlcus. Amphipholis kochii. Amphip}h olis japonica . ^^2 ABT. 2. — H. MATSUMOTO Ampliqyliolis sohrina. Am])hiura dlgitula. Amiohiura tracliydisca. Amphmra microdiscus. AmjyJmira euopla. Amphiura pycnostoma. Amphiura acacia. Amp)hmra iridoides. Amphiura micraspis. Amphiura œstuarii. Amphiura vadicola. Amp)hiura ecnomiotata. Ophiothrix jyanchyendyta. Ophiothrix marenzelleri. Ophiothrix macrohrachia. Ophiothrix eusteira. Ophiothrix stabilis. Ophiomaza hanehoi. Astrophiura kaicamiirai. Aspidophiura luatasei. Sfegojyhiiira striata. Stegojjhiura vivipara. Stegophiura scidj^ta. Stegophiura sterea. Stegoph iura hr achy actis. Stegophiura sladeni. Ainph ioph iura œdiplax. AmjMoph iura pompophora. Amphioph iura penichra. Amphiophiura megapoma. MONOGRAPH OF JAPANESE OPHIUROIDEA. 363 Amphioph Iura lapidaria. Homalophiura clasta. Oph'mra o'ôplax. Ophiura micracantha. Ophiura paucisquama. Ophiura monostœcha. Ophiura albata. Ophiura calyptolepis. Ophiomusium granosum. Ophiomiisium irychnum. Ophiosonella elevata. Oph iozoneUa platydisca. Ophiozonella polyplax. Oph io^o7iella longisplna. Ophioplocus japonicus. Ophioleuce charischema. Ophioleuce hremspinum. Ophiotroehus longispinus. Ophiiiroconis monolepis. Bathypectiniira gotoi. Pectinura aiichisfa. OpkiarachneUa megalaspis. Ophioplax lamellosa. Ophiodoris per/cafe. Ophionereis sinensis. Ophiocrasis dictydisca. Ophiocrasis marldanneri. 140 species: 60^. The three groups of Cirumpolar, Bering- American and Bering species are evidently referable to the Arctic fauna, and form all 364 AKT. 2. — H. MATSUMOTO : together S% of the total number of Japanese species ; while the three groups of Intertropical, Indo-Pacific American and Indo-Pacific species unquestionably represent the tropical elements, and form all together 28X of the total number. Among the Arctic species, three have the southern limit in the Okhotsk Sea, three in the Sea of Japan northward of Sado, five in the Sea of Japan off southern Korea, and four in the Eastern Sea ; and along the Pacific coast of Honshu, two in the Sagami Sea and Uraga Channel, four in the Yenshu Sea and Suruga Gulf, and three in the Kumano Sea, off Kii. The Arctic species, which occur in the Eastern Sea, appear to range through the Sea of Japan, but not through the Pacific coast of Honshu. Among the tropical species, one has the northern limit in the Strait of Formosa, two at Pinnacle Is., seven at Okinawa, one at Benin Is., three in the Colnett Strait and Kagoshima Gulf, one at Natsui, two in Tanabé Bay and Kumano Sea, one in the Yen- shu Sea, one at Eno-ura, twenty-five in the Sagami Sea and Misaki region, one at Kominato, one off Bôshû, one in the Pacific Ocean far off Honshu, one off Kinkwa-san, one in British Columbia and three in the Bering Sea ; and by way of the northern part of the Eastern Sea to the Sea of Japan, six in the Eastern Sea, five in Korea Strait and the Sea of Japan off southern Korea, one off Echizen, one off Echigo, one in Hakodate Bay, two off Shiribeshi and one at Wladiostok. Now, neglecting here the very hardy species, which have their northern limit in British Columbia and in the Bering Sea, the zone of mingling of the Arctic and tropical species extends from off Kinkwasan to the Kumano Sea, on the Pacific side, and from off Shiribeshi and Wladiostok to the northern part of the Eastern Sea, on the side of the Sea of Japan. Thus, the geographical MONOGIÎAPH or JAPANESE OPHIUKOmEA. 365 boundary between the two elements is better marked on the Pacific side than on that of the Sea of Japan. The Hraits of the zone of contact are approximately 33^. and 38J. parallels on the Pacific side, and 80. and 43. parallels on the side of the Sea of Japan. The axis of the zone of contact is very nearly 36. parallel, which has been stated by Claek to be the line of contact between his Bering and Honshu faunae. The two groups of Honshu- American and Honshu species are evidently heterogeneous. For example, Amphlpholis imgetana, Amphiura carchara, Gorgonocephalus tuberosus, G. dolicliodactylus, Ophiolehcs hrachijgnatha, 0. asaphes, 0. tuherosa, Ophiosemnotes œdldisca, Opliiopliolis mirabilis, 0, bracliyactis, Stegophiura striata, St. vivipara, St. sculpta, St. sterea, St. brachyactis, St. sladeni, &c, may be derivatives of the Arctic fauna, while the rest are largely derivatives of the tropical, especially Indo-Pacific, fauna. Claek has recognised scarcely a dozen species which occur in both the East Indian and Japanese regions. As a result of the present study, I recognise sixty-one species, or more than one-fourth of the total number of Japanese species, that occur also in the East Indian region. On the other hand, the boundary between Claek's Bering and Honshu fauna) well coin- cides with that between my Arctic, chiefly Bering, and tropical, chiefly Indo-Pacific or East Indian, elements. Thus, I am led to look upon the Honshu fauna not as a perfectly distinct one, but to be a terminal section of the Indo-Pacific, or East Indian, fauna. The known northern limits of several more typical, httoral Indo-Pacific species are shown in the accompanying chart. 366 ART. 2. — H. MATSUMOTO : 120 va 124 126 128 130 132 134 136 1. Pinnacle Is. — Ophiotlmx hirsuta, Ophiomastix annulosa. 2. Okinawa Is. — Euryale aspera, Ophiothrix dbtusa, Ophiomastix lutkeni, Ophiarthrum elegans, OpHolepis annulosa, OpMoplocus imbricatus, Opldarachrut incrassata, 3. Bon in Is. — Ophiarachna incrassata. 4. KagosMma Btiy.— Ophiocoma scolopendrtna. 5. Natsui. — Ophiocoina hrevipes. G. Tanabé. — Triehaster elegans. 7. Eno-ura. — Ophiarachnella inf emails. 8. Misaki & Sagarni Sea. — Ophiothrix nereidina, Ophiadis modesta, Ophiocentrus verticillatus, OphiaracJmella gonjonia, Ophiomastix mixta. 9. Kominato. — Ophiothrix longipeda. MON0GR.y:'H OF JAPANESE OPHIUROIDEA. 367 Phylogeny of the Ophiuroidea. It is very evident, that the forms with vertically coiled arms are not archetypal, because th(^y have very solid and always undivided vortebrœ, as mentioned in the introduction, while the ophiuran vertebrae are palœontologically, as well as ontogenetically, proved to be primarily divided. Thus, the OpMohyrsince, Trichas- tericlœ, GorgonocephaUdœ, Ophiacanthidœ pars, Hemieuryalidœ, Ophiaclince pars, and Ophiotrichidœ pars are eliminated from primitive ophiurans. The forms with well developed lateral wings of the oral fixâmes are evidently highly specialised. Thus, the Amphiuridce, Ophiotrichidœ, Ophlolepidlnœ pars, Ophionereidince and Ophiocomidce are eliminated from primitive ophiurans. The remaining forms are the Ophiomyxlnce, Ophiocanthidce pars, Amphilepididce, Ophiolepididœ pars, Ophioleucidce, OphiodermaUdce and Ophiochitomnœ. Indeed, I have been able to find more or less divided vertebrae in certain representatives of the Amphilepididce, Ophioleindldœ and Ophiodermatidce, as well as of the Ophiomyxince and Ophiacanthidœ. Though it is my purpose to discuss the results of a study of Palaeozoic ophiurans in a future paper, I will here enumerate some of the more important structures of Palaeozoic Myophluroida, as bearing on the question before us. 1. Disk covered with delicate scales or by a naked skin, without distinct primaries. 2. Kadial shields absent. 3. Genital plates and scales absent. 4. Oral shields absent. 5. Adorai shields not very distinctly speciaUsed û'om the lateral arm plates. 368 ART. 2. — H. MATSUMOTO : 6. Oral plates and frames long and slender. 7. Distinct creases probably present between the interbrachial ventral surfaces and arm bases. 8. Dorsal arm plates entirely absent, or present only in a few basal joints ; the dorsal side of the arms therefore largely unprotected. 9. Lateral arm plates with prominent spine ridges, which extend to the ventral side of the arm ; those of the two sides not meeting above or below, except in the very distal arm joints. 10. Ventral arm plates higher in position than the lower borders of the latei-al arm plates, so that the arm is longitudinally grooved ventrally. I believe that, the Palaeozoic Myophiuroida are the stock from which the recent ophiurans have been directly derived, because they show no trace of peculiar specialisation and are fairly intermediate in their organisation as a whole between the Œgophiuroida and recent ophiurans. If this view be right, then the most archetypal group of recent ophiurans must be looked for among those forms which have the strongest resemblances to the Palaeozoic Myophiii- roida. Certain genera of the Ophiolepididce, which are frequently said to be primitive, appear to me to be far from being archetypal, though they may evidently be paodomorphic. These genera in the mature stage are in every feature similar to very young stages of other ophiurans, having the disk covered only with the primaries and radial shields, and the arms covered chiefly by the lateral arm plates. But, the existence of such forms in the early ages of the history of the Ophkiroidea is entirely unproved. As to the biogen- etic law, VON Baer's view appears to me to be more consonant to facts than F. Müllee's. The term "primitive" may mean MONOGEArn OF JAPANESE OPHIUKOIDEA. 369 both " archetypal " and " embryonal " ; but the two must in my opinion be strictly distinguished. In fact, the genera in question of the Ophiolcpklidce are, in my opinion, embryonal, but not archetypal. On the contrary, the Ophiomyxince and certain genera of the Ophiacanthidce with only horizontally flexible arms, appear to me to be fairly archetypal, being similar to the Palaeozoic Myophmvoida in many structures, the presence of the genital plates, genital scales and oral shields, and sometimes also of the radial shields being the principal differences. As to the radial shields, I look upon them as a modification of the marginal disk scales, often present in both the Palaeozoic and recent ophiurans, which have secon- darily become articulated with the genital plates where these have developed ; and a very primitive condition of the radial shields. may in my opinion be seen in Ophiostiba, Ophiohyalus, as well as in young specimens of Ophiomyxa, &c. This view is in accord with Lyman 's^\ according to which the radial shields are not special plates but are homologous with the other disk scales, and by no means the first to appear. The ventral arm plates of many genera of the Ophiomyxince, as Ophiosciasma, Ophiocynodus, Ophiosysygus, Ophiolepioplax, Opkio- styraciwn, Opliioliyalus, Ophiomyxa, Ophiodera, Ophiohymen, &c., and some of the Ophiacanthidce, as Ophiomyces, Ophiologimus, &c. are long and narrow, being in contact with, or slightly separated by a naked space from, each other, at least in the proximal arm joints. A similar condition may be observed also in certain repre- sentatives of Palaeozoic Myophiuroida. In many genera of the Ophiacanthidce, the ventral arm plates are very small, short and 1) Eep. Challenger, V, 1882. p. 157. 370 ART. 2. — H. MATSÜMOTO : widely separated from one another by the lateral arm plates, which are distinctly in contact with each other in the ventral median line. This condition is also observed in the distal arm joints of the Ophiomyxince ; so that, following the law of localised stages, it is evidently more distinctly embryonal than that in which the lateral arm plates are separated by the ventral arm plates. The Œgophiuroida had exposed ambulacral grooves ; and this groove became covered over by the ventral arm plates for the first time in the Palaßozoic Myophiuroida ; so that, the meeting of the lateral arm plates in the ventral median line can not be a primary condition. Moreover, the assumption that the meeting of the lateral arm plates in the ventral median line is a primary condition, makes it very difficult to assign a probable cause of the acquisition of the ventral arm plates. I therefore think that, the separated condition of the lateral arm plates in the ventral median line is an archetypal but not embryonal feature, while the meeting of those plates is an embryonal but not archetypal feature. In the Ophiomyxince, the lateral arm plates of the two sides are widely separated from each other on the dorsal side, while in the majority of the Ophiacauthidce, they are well in contact. The second condition is what occurs also in the very distal arm joints of the Ophiomyxince, as well as of many Palaeozoic ophiurans ; so that, following the law of localised stages, it is evidently more distinctly embryonal than the first condition. But, the assumption that the meeting of the lateral arm plates in the dorsal median line is a primary condition makes it very difficult to assign the neces- sity for the acquisition of the dorsal arm plates. I therefore think that, the meeting of the lateral arm. plates in the dorsal median line is an embryonal but not archetypal condition, while the MONOGRAPH OF JAPANESE OPHIUROIDEA. 371 separation of the same is an archetypal but not embryonal con- dition. The interpretation put forth in this and the last paragraph is well in accord with the biogenetic law as formulated by von Baer but not as formulated by F. Müller. In Ophiolepfoplax and Ophiohyalus, the dorsal arm plates are present but very rudimentary, thin and hyaline. A similar con- dition may be observed also in the distal arm joints of Ophiomyxa, as well as Ophiodera anisacmitha, though the plates are in these forms divided into several secondary plates in the proximal arm joints. I imagine that, these dorsal arm plates may represent an archetypal condition. As to the divided vertebrae, those of tlie Ophiomyxince appear to me to be more archetypal than tliose of the Ophîacanthidœ, the former reminding us of those of such a Palaeozoic form as Ophiur'ina. Upon the basis of these considerations, I imagine that the Ophiomyxince are a step more archetypal than the Ophiacanthidce. Those genera of the Ophiomyxince, in which the second oral tentacle pores open outside the oral slits, have usually very slender adorai shields and very small oral shields, reminding us of the Palaeozoic Myophiurolda, which are characterised by the second oral tentacle pores opening outside the oral slits, by the adorai shields not being well differentiated from, but almost similar in shape to, the ordinary lateral arm plates, and by the absence of the oral shields. That the second oral tentacle pores open outside the oral slits is a character, which appears to be at once archetypal and embryonal. The majority of those genera of the Ophiacanthidce, in which the arms are only horizontally flexible, have more or less well developed dorsal and ventral arm plates, which separate the lateral 372 ART. 2, — H. MATSCJMOTO : arm plates of the two sides more or less completely. These genera strongly resemble the Ophiomyxince, especially Ophioscolex, save in the presence of well developed dorsal arm plates. I look npon them to be a most archetypal, though not embryonal, group of the Ophiacanthidœ. I of course believe that the gap between the Ophiomyxmcc and Ophiacanthidœ is rather not very important. Thus, the phylogenetic origin of the recent ophiurans may be shown somewhat as follows. Ophiacantliidsei Ophiomysinse Palaeozoic Myophiuroida Now, I will trace the Ophiomyxince-lino further. The Ophio- byr since are evidently a step more advanced from the Ophiomyxince toward the Trichasteridœ and Gorgonocejjhalidce, having very com- pact oral skeleton and more or less short, stout, rather discoidal vertebras with streptospondyline articulation. The only essential distinction of the Ophiohyrsince from the Trichasteridœ and Gor-. gonocephalidce arises from the fact, that the vertebrae of the sub- family in question are less discoidal and have less well developed upper and better developed lower muscular fossae, as compared with those of the two last mentioned families. Among the Trichasteridœ, the Aster onychince appear to me to be most archetypal, because the disk is large, the arms are com- paratively slender, the lower muscular fossae of the vertebrae are relatively little reduced and the arm spines are numerous, instead of being reduced to two in number, thus more or less reminding MONOGRAPH OF JAPANESE Oi'HIUROIDEA. 378 US of the Opliiohyrslnœ. Tlie Trlchasterince are, in my opinion, nearer to the Aster onycJiince than the Aster oschemat'mce are to the same, because the ventral arm plates of both the Asteronychinœ and I'richasterince are rather well developed and in contact with one another, entirely separating the lateral arm plates, while those of the Asteroschematince are very much reduced in size and separated from one another by the lateral arm plates, which meet in the ventral median line. This view almost coincides with that of MoRTENSEN, wlio has pointed out the affinity of Aster onyx to Euryale. The Asteroschematince are evidently descended from the simple-armed group of the Trlchasterince. Whether the Gorgonocephalidœ are descended from the Ophiohyrsince or from the Aster onychince is rather hard to decide, though it is very evident, that the Gorgonocephaliclce have no direct relation to the Trlchasterince and Asteroschematince. The minute hooks of the arm annuli stand in an intimate relation with the arm spines, as may be clearly witnessed in the very distal arm joints, where the arm spines are hook -shaped and the upper ones are smaller and show a tendency to sliift their positions alternately so as to form a double row. These upper arm spines are evidently the rudiments of the arm annuli. As already mentioned, the lateral arm plates are approximated dorsally in embryonal stages ; and such a dorsal approximation occurs, in the present case, in the arm spines. So I imagine that the arm annuli are due to the persistence of an embryonal feature of the arm spines. If this interpretation be right, then the arm annuli of the Gorgonocephaliclce remind us of the arm, spines of such genera as Ophiobrachion and Asteronyx. Though Opkiohrachion is referred to the Ophiohyrsince of the Ophiomyxldce and Asteronyx to the Aster onychince of the Trlchasterldce, the difference between 374 ART. 2. — H. MATSUMOTO : the two subfamilies is by no means very great. So that, it may not be very unreasonable to look upon the TricJiasteridce and Gorgonocephalidce as having a blood relation att he very begin- ning. Within the limit of the simple-armed, genera of the Gorgono- cephalidce, the Gorgonocephcdinœ are evidently more archetypal than the Astrotomhue, because the former have less specialised and less localised oral and dental papillae and less powerful oral angles than the latter, and the latter have well developed supple- mentary plates in the spaces between the oral angles and the interbrachial ventral surfaces. Another point to be noted is that the dividing of the arms occurred independently in three groups, viz. the TricJiasterinœ, Gorgono2ephalince and Astrotominœ, an example of parallelism due to certain biological conditions. The interrelationships between the members of the Ophio- myxinœ-\m.Q, or Phrynophiurida, may be shown as follows. Gorgonocephalin œ •with dividerl a,rms Astrotominœ with divided arms I Astrotominae with simple arms \ Gorgonocepliali ase with simple arms Tricliasteringe with divided arms Asterosclieraatinse Trichas teriuae with simple arms Asteronychinre Opbiobyrsinœ Opliiomyxiuse MONOGRAPH OF JAPANESE OPHroROEDEA. 375 I will now proceed to the consideration of the Ophiacanthidce- line. As already mentioned, the genera or species with only horizontally flexible arms are a step more archetypal than those with more or less vertically coiled arms ; but the distinction of these two groups can not be very profound, since certain genera include species of both types. As far as the genera examined by myself are concerned, Ophiologimus and OphioUmna are typical representatives of the forms with only horizontally flexible arms, and OphioJehes, Ophiosemnotes and Oph'iochondrella of those with more or less vertically coiled arms. Ophiologimus, as probably also Ophiotoma, OpJiioblenna, Ophiomyces, Ophiocimhium, Ampliipsila, &c. may represent a most archetypal group of the Ophiacanthidce ^ being very close to the Ophiomyxince on the one hand and to the Amphilepididce on the other. Microphiura may be very embryonal, having the oral papillae completely soldered together and the lateral arm plates extremely well developed. In a very small and young specimen, of which the genus is indeterminable, I have observed that every essential structure is similar to that of Microphiura, the oral papillœ being completely soldered together, the genital slits invisible, the second oral tentacle pore opening entirely outside the oral slits, the tentacle pores provided with a crescent-shaped scale, and the lateral arm plates almost completely covering the entire surface of the arm. I therefore consider Microphiura to be probably a pa3domorphic form, granting that the spscimens described by MoRTENSEN are sexuaUy mature ; it is in my opinion not truly archetypal. I have already mentioned that the very good development of the lateral arm plates, which separate the dorsal and ventral arm plates from each other, is an embryonal feature. Following this principle, the majority of the Ophiacanthidce are embryonal in the 376 AET. 2. — H. MATSÜMOTO; character of the arm plates. The Palaeozoic ophiurans are usually very large, one specimen of Palaeozoic Myophiuroida at hand being about as large as a very large specimen of Ophiomyxa ßacclda, and another as large as, or larger than, a very large specimen of Ophiarachna incrassata ; while the Ophiacantliidce are usually very small, and their arms so slender as to equal in thickness only the very distal part of the arms of the Palaeozoic Myophiuroida. I imagine that the very small size of the recent Ophiacantliidce is intimately correlated to their embryonal characters. The Hemieuryalidce are, in my opinion, evidently a terminal group of the line of speciahsation represented at the base by Ophiolehes, Ophiosemnotes, Ophiochondrella, &c., which are character- ised by the pronounced vertical coiling of the arms. Among them, the Ophiochnodrinœ are very close to the Ophiolehes -gvow^ and more archetypal than the Hemieuryalince. The interrelationships of the Lœmophiuridà may be shown as follows. Hemieiiryalinse OpliiocliondriuBe I I Ophiacantliidse with more or less vertically coiled arms. OpliinciLiitlnd80 with only horizontally flexible arms. From the Lœmophiuridà upwards, there are, in my opinion, two distinct lines of specialisation : one, forming the Gnathophiurida^ is characterised essentially by the radial shield and genital plate articulating with each other by means of one large, conspicuous socket of the former and one large, ball-like condyle of the latter ; MONOGRAPH OF JAPANESE OPHIUEOroEA. 377 and the other, forming the Chllophiurlda, is characterised essen- tially by the radial shield and genital plate articnlating with each other by means of two condyles and one pit of either plate. Among the Gnathophiurida, the Ampliilepidldce are evidently most archetypal, being closely allied to the Ophiacanthidce with only horizontally flexible arms, in the very large peristomial plates, in the not very stout oral frames without well developed lateral wings, in the very long and slender oral plates, in the genital plates being not fixed to, but free from, the basal vertebrae, and in the often incompletely divided vertebrae. The Amphiuridœ and OphiotricJiidce are characterised by the very small peristomial plates, by the very stout oral frames with well developed lateral wings, by the stout oral and dental plates, by the very stout, quadrangular teeth and by the genital plates being firmly fixed to the basal vertebrae. Among them, the Ophiactinlnce are evidently nearest to the Amphilcpididce, because neither have paired infradental papillœ or dental papillae. In my opinion, the paired infradental papillae of the Amphiu- rince and the dental papillae of the OphiotricJiidce were acquired as supplementary organs for mastication after the teeth had become quadrangular and very stout. The Ampliiurinœ might be looked upon as ancestors of the Ophiotrichidce, if the paired infradental papillae and dental papillae could be proved to be homologous. But according to Clark,' ^ the paired infradental papillae are genuine oral papillae, while the dental papillae are modifications of the teeth. Hence it appears that the two kinds of papillae just mentioned are merely analogous organs adapted to a similar function. The Amphiurince and Ophiotrichidce are then parallel 1) Growth-Changes in Brittle Stars. Publication No. 182 of the Carnegie Institution, Washington, 1914. 378 ART. 2. — H. MATSUMOTO : groups and not directly related to each other, not to mention that the Ophîotrichidœ are the most highly specialised of the Gnatho- phiurida. It appears to me to be noticeable that in those forms, in which the masticatory apparatus is well developed, the oral papillae are liable to reduction. In Ophiacfis, Ophiopiis, Hemipholis, the Am- pliiodia-gvoxï^, &c., the papilla3 which arise from the adorai shields are absent. In OpMopliolis and the Ampliiura-^oxï^, the oral plates are almost free of papillae. In Ophiopsila {Ophiocomidœ), the outer oral papillae are strongly reduced. And finally, the Ophîo- trichidœ entirely lack the oral papillae. This fact probably has a certain biological meaning. The interrelationships of the Gnatliophiurida may be shown approximately as follows. Ophiotrichidse / . / - Ampliinrinœ / Ophiactiniuas Amphilepididse Ophiacanthidse ■u-ith only horizontally flexible arms. The Chilophiurida are very extensive and divergent, and it is very difficult to make out their interrelationships clearly. Certain genera of the Ophîolepididœ, e.g. Astrophiura, Ophiomisidium, Ophio- mastus, &c., and Ophiomuslum, are frequently looked upon by certain authors as very primitive. I can agree with these authors, MONOGRArH OF JAPANESE OPHIÜROIDEA, 379 if by " primitive " they mean " embryonal " in the strict sense. But if they regard the genera in question as archetypal, I must dissent. My reasons for doing so may be stated as follows. 1. The disk of these forms is entirely or largely covered ■with the primaries and radial shields, a condition observed in very young and small individuals of other ophiurans. The disk structure of these forms is due to the simple growth of the plates found in a very young stage. 2. In these forms, the genital slits are either entirely invisi- ble or exceedingly small. The genital slits are also entirely in- visible in very young and small individuals of other ophiurans. 3. The oral papillsß of these forms are usually fused together to form a single piece. This feature probably represents the stage previous to the division of a common rudiment into individual papillae. 4. The arms of these forms are extremely short, consisting of only a small number of joints, and are covered chiefly by the lateral arm plates. This condition is also observed in very young and small individuals of other ophiurans. 5. In these forms, the tentacle pores are limited to a few basal arm joints. It appears to me very difficult to look upon this condition as archetypal ; it may however be embryonal. 6. The wings of the vertebrae are in these forms exceedingly thick even in the very basal arm joints. This condition may be derived from what is observed in a very young stage by simple growth of the vertebrae without change in the relative size of their constituent parts. Thus, the genera in question appear to me to be paedomorphic in almost all structures ; so I look upon them as neotenic forms. That they can not be truly archetypal, and that embryonal i.e. 380 ART. 2. — H. MATSUMOTO : paedomorphic characters are not always archetypal, is clear from what I have said already. In my opinion, the Opliiarachnince and Ophiochltonince are the most archetypal of the Chilopliiurlda, some of them being so near to the Ophiacanthidce as to be distinguished from the latter only with great difficulty. Indeed Opliiuroconis and Ophiurochœta have hitherto been confused with Ophiolimna of the Ophiacanthidce by several authors, and Ophiochiton was considered by Verkill to belong to the Ophiacanthidce, Moreover, I have been able to find out incompletely divided vertebrae in certain species of OphiuroGonis and Ophiurodon. A next ally of the Ophiarachnince is, of course, the Ophioder- matinœ, which are distinguished from the former by the very short arm spines. Judging fi'om the degree of calcification of the vertebrae, the Ophiarachnince ■ are evidently more archetypal than the Ophiodermatinœ. The Ophiochitonince lead to the Ophionei'ei- dince, which are however a step more specialised than the former in having a well developed masticatory apparatus. Ophiodoris is evidently intermediate between the Ophiochitoninœ and the other Ophionereîdinœ. The Ophionereidince are intermediate between the Ophiochitoninœ and Ophiocomidce, so far as the oral skeleton is concerned, and might fairly be looked upon as the direct ancestor of the last mentioned family, if the dorsal surface of the vertebrae were not so strongly notched at the inner end. This last character of the vertebrae of the Ophionereidince, as well as probably of the Ophiotrichidœ, appears to me to be correlated to the very high flexibility of the arms. The same character is not however equally well developed in all the Ophionereidince, but only very feebly in Ophiodoris, wliich is unique among the Ophionereidiiue in being destitute of supplementary dorsal arm plates. A parallel character MONOGRAPH OF JAPANESE OPHIUKOIDEA. 381 is found in the Ophlocomidœ, in which the dorsal surface of the vertebraß is not distinctly notched at the inner end and the dorsal arm plates are entirely free of supplementary plates. Therefore, it appears to me not very improbable that the Opliionereidince and Ophiocomidce may finally merge into a common stock at the very base. I have observed that very young specimens of a certain species of Ophiocoma have a squamated disk, which is almost free of granules, and arms which in ventral view remind us of those of Ophioplax and Ophioneretdmœ. As already pointed out, the oral papillae are liable to be reduced when the masticatory apparatus is very w^ell developed. This tendency is also observable in the Ophiopsilince {Opliiocomîdœ). Moreover, the Opiuopsilinœ appear to me to be slightly more specialised than the Ophiocomhiœ, because the radial shields are fairly bar-like, the outer ends of the genital plates more or less approximated to each other at the dorsal side of the arm base and the tentacle scales very peculiar in shape. I imagine that the Ophiopsilinœ represent a last phase of the Ophiocommœ. As suggested by Kœhler, Ophiochiton appears to me to re- semble much such genera as Ophlozona and Ophiozonella of the Ophiolepidinœ. This resemblance might be of a superficial nature, if the difference of the arm spines being erect or appressed were of primary importance. But this diifference is observable even in the proximal and distal parts of a single arm in many genera. I therefore agree with Clark, who looks upon the diiference of the arm spines in question not to be of primary importance. Then, the relation of Ophiochiton to Ophiozona, Ophiozonella, &c. cor- responds to that of the Ophiarachnince to the Ophiodermatlnœ. Among the Ophiolepidinœ, Ophioceramis has a very well developed masticatory apparatus. Thus, we see that the mastica- 382 ART. 2.^H. MATSUMOTO : tory apparatus is very well developed more or less independently in three lines, one of wliich is represented by the AmpJiiuridce and Ophiotrichidce, the second by the Ophionereidinœ and Ophiocomîdœ, and the third by Ophioceramis. As I have already explained, OpMomusium is the most embryonal, i.e. paedomorphic form of the OpMolepidince. In my opinion, the Ophiomastinœ present an almost unbroken series of psedomorphism, of which the terminal members are such genera as Astrophiura, Ophiophycis, Ophiomisldium, Ophiotypa, Ophiomastus, Anthophiura, Aspidophmra, Ophlopyrgus, &c. The grounds for this opinion have already been stated. If I am right, the Ophiomastinœ, as well as Ophiomusium, must have been derived from the other OpMolepidince by neoteny. The Ophioleucidce superficially resemble the Ophiodermatidœ, but their true allies are, in my opinion, the Ophiolepidince. The very thick and solid squamation under the superficial granules of the disk, the usually internally joined radial shields, the rhomboidal or lyre-shaped oral shields, the peculiarly long adorai shields, the comparatively long oral plates, the articulation of the genital plate and scale at their outer ends, the flattened arms, the very few — usually only two — arm spines, &c. are characters usually observed also in Opliiura and its allies. In certain genera of the Ophioleucidce, the second oral tentacle pores show a tendency to be displaced outside the oral slits. Ophiotrochus, which has partially soldered oral papillae and very rudimentary dorsal and ventral arm plates, may be the most embryonal, i.e. paedomorphic form of the present family. It is theoretically to be expected that the very embryonal forms of various families, e.g. Microphiura of the Ophi acanthi dee, Astrophiura, Ophiophycis, Ophiomisidium, Ophiotypa, Ophiomastus, MONOGRAPH OF JAPANESE OPHIUBOIDEA. 383 AnthopJdura, Aspidophiiwa, Ophiopyrgus^ &c. of the Ophiomastince, Ophiomusiiun of the Ophlolepidlnce and Ophiotrochus of the Ophio- leucldce, apparently resemble one another in some characters ; bnt these resemblances have in my opinion no significance for the phylogeny. The interrelationsliips of the Chilophmrida may be shown as follows. Ophioleucidae ■with very rudimentary dorsal arm i^lates Ophioleucidse with well developed dorsal arm plates Ophiomastinse Ophiolepidinse with only a few tentacles Opliiolepidiuse with well developed masticatory apparatus Ophiolepidinse pars OpliiopsilinsB Opliiocomins9 Ophiodermatinae Ophiochitoninae Ophiaraclminae Ophiacanthidsö with only horizontally flexible arms 384 ART. 2. — H. MATSUMOTO : Postscript. Recently I received Clakk's " Catalogue of Recent Oplainrans," Mem. Mns. Comp. ZooL, Vol. XXV, No. 4. As this paper of mine was then in press, no reference is made to it in the Hteratnre of the species. My opinions or notices upon several parts of his monograph have been added as foot-notes. General Explanation of Plates. PI. I, figs. 1-9 illustrate the internal stractures of the Ophmnyxina:. The peristomial plates are double in Ophiomyxa australis (figs. 4 & 7) and Ophiostiha hidekii (fig. 9), one overlapping the other in the latter species ; triple in Ophiodera anisacnntha (fig. 1); double or triple in Ophio- hgalus gotoi (fig. 8). It is noteworthy, that the peristomial plates show a tendency to be simple in younger stages, as may be seen in fig. 4, illustrating a younger siDecimen of Ophiomyxi misiralis. The double or triple (at least in adult), thin, delicate peristomial plates, and not very stout oral plates and frames are characteristics of the Ophiomyxinœ in. contrast to tbe Opldo- byrsina', Tiichastervdœ and Gorgonocephalîdœ. The articulation of the genital plate and radial shield is illustrated in fig. 1 {Ophiodera anisacaiithu) and fig. 4 [Ophiomyxa australis); it is very simjjle, without any particular condyle and socket, a common character of the Phrynophiurlda and Lcemophiurida. In the last two illusrations, the genital scales are seen to articulate with the genital i^lates at some distance away from the outer ends of the latter, a common character of the OpMomyxidœ and Asieronychina: in contrast to the Trlchasterbia:, Asteroschematinœ and Gorgonocephalîdœ ; and the dorsal surface of the vertebrre (excejit the very basal ones) is rhom- boidal and the wings of the same very thin on the sides, a characteristic of the Ophiomyxinœ in contrast to the Ophiohyrsina;, Trichasteridœ and Gorgonocephalidce. The articular surfaces of the vertebrœ are illustrated in figs. 2 & 3 (Ophiodtra anisacantha) and 5 & 6 {Ophiomyxa australis); the articulation is typically zygospondyline, the umbo and knobs on the inner articular surface, and the shoulder and peg on the outer, being all well developed, a character of the Ophiomyxinoi in contrast to the Ophiohyrsina-, Trichasterbue and Gorgonocephalîdœ; the great size of both the iipper and lower mi^scular fossœ is a character common to the Ophiomyxida' and the other " common " ophiurans in contrast to the Trichasterido' and Gorgonocephalidœ. PI. I, figs. 10 & 11 and PI. II, fig. 1 ulustrate a vertebra of Ophiosmilax mirahilis, a repre- sentative of the Ophidbyrsina'. The articulation is typically streptospondylino or saddle- shaped _ a common character of the Ophiohyrsinœ, Trichasteridœ and Gorgonocephalidœ in contrast to the Ophiomyxina:. The upper and lower muscular fossœ are both large, a character of the Ophio- MONOGRAPH OF JAPANESE OPHIUEOIDEA. 385 rnyxidtv in contrast to the Trichasteridœ and Gorgonocejjfutlvlo: The vertebra is moderately short and stout (PI. II, fig. 1), being shorter and stouter than that of the Ophîomyxinœ bitt not so very short and stoixt as in the Trichasteridœ and Gorgonocephalida; a characteristic of the Ophidbyrsino'. The wing of the vertebra is very thick laterally as well as dorsally, a common character of the Ophidbyrsinœ, Trichusieridœ and Gorgonocephalida'. As the specimen, from which the illastration has been made, is very small and young, the characteristic shape of the vertebra is not yet fully developed ; the vertebrre near the disk of full-grown specimens of the Ophidbyrsinœ are much more stout and more distinctly discoidal. PI. I, figs. 12 & 13 show the two forms of arm spines of OphiosmUax mirabilis, both being converted into c impound hooks. It is noteworthy, that the arm spines of the majority — presumably all — of the I'hryno- phiurida are always book-shaped in their embryonal stages, as observed in the distal arm joints. PI. I, figs. 14-16 illustrate the internal structures of Asteronyx loveni. a representative of the Aster onychînœ. The peristomial plates are simple, stout and firmly soldered to the oral frames, which are also very stout, common characters of the Ophiohyrsina\ Trichasteri/Jœ and Gorgono- cephalida; in contrast to the Ophio7vyxina.\ The veiy basal vertebrre are not notably shorter than the outer; all the vertebra; are short, stout and discoidal; and the vertebral articulation is typically streptospocdyline or saddle-shaped ; all th( se characters are common to the Ophio- tyrsina?. Tricha stirida; and Gorgonocephalida'. The upper muscular fossae are fairly large and the lower rather small ; this character is rather intei mediate between those of the Opîdoiiiyxidc' and of the other Trichasterida', for the upper and lower muscular fossa; are subequal in size in the former and very unequal in the latter. The relatively small inequality between the ujjper and lower fossa' here illustrated may partially be due to the youngness of the specimen. The genital jjlates are high in position relatively to the basal vertebra; ; this character also approach- es to that of the Ophiomyxidœ. The genital scales articulate with the genital plates near the inner ends of the latter, a characteristic of the Asieronychkuv in contrast to the Trlchasterina?, Asteroschemalina: and Gorgonocephalida:. The articulation of the genital plate and radial shield is very simple, without any particular condyle and socket, a common character of the Fhryno- phiurida and La'mopjhiurida. PI. n, figs. 7 it 8 illustrate the skeletal structure of Trichaster elegans, a representative of the Trichasterina'. The oral plates are very stout, a common character of the Ophiobyrsina', Trichasteridœ and Gorgonocephalida^ The adorai shields are also large and stout, a character of the Ophiomyxidœ and Trichasterida; in contrast to the Gorgonocepjhulida' ; a jjeculiar feature is that they are divided into inner and outer halves. The lateral arm plates of the two sides are separated from each other by the ventral arm plates, a character of the Asteronychiva' and Trichasterina; in contrast to the Ast roschematina;. The ventral arm plates are divided into two or three secondary plates, a peculiar feature. The interbrachial area is very small and the genital plates and scales are very closely set side by side ; this is a generic character. That the genital scales do not lie far inward from the oiüer ends of the genital plates is a common character of the Trichasterina', Asteroschemathxa: and Gorgonocephalida: in contrast to the 386 ART. 2. — H. MATSÜMOTO : Aster onychince. Text -fig. 8 h-ü (p. 38) muy also be referred to here. The upper imiscular fossîe of the vertebra are much larger than the lower, a common character of the Trichasteridcp. and Oorgonocephnlidœ ; each dorsal arm plate being represented by a pair of triple rows of nodule- like secondary plates is a generic character ; that each dorsal arm plate is represented by a pair of secondary plates or of groups of secondary plates in the basal arm joints containing the generative glands, is a common character of Astroceras, Trichasler and Asteroschema., PI. II, figs. 2-6 iUustrate the internal structure of Asteroschema japonicum, a representative of the Asteroschemalina\ The very stout peristomial plates which are firmly soldered to the also very stout oral frames, the short and stout and markedly discoidal vertebr;«, the strepto- spondyline vertebral articulation, the markedly unequal tipper and lower muscular fossne of the vertebrae, the very simple articulation of the genital plate and radial shield, and the position of the genital scales in relation to the genital plates, are characters which have been already explained. The lateral arm plates are transversely bar-like, and those of the two sides are in contact with each other in the ventral median line, thus separating the very small ventral arm plates from one another, a characteristic of the Aster oschematinœ. Each dorsal arm plate is represented by a pair of bar-like rows of secondary plates ; this is a generic character, though the fact that each dorsal iilate is represented by a pair or groiip of secondary plates is a common character of Astroceras, Trichaster and Asteroschema, as already explained. PI. II, figs. 9 & 10 illustrate the internal structure of Astrotomn sobrina, a representative of the Astrotomina; {Gorgonocephalida;). The common characters of the Trichdsterldœ and Gorgono- cepalidce are sufficiently exf)lained in the text. The basal vortebrre are much narrower and smaller than those beyond, a characteristic of the Astrotominœ in contrast to the Gorgonocephalinœ. They are covered over by the muscles between them and the genital plate, also a characteristic of the Astrotominœ. The adorai shields are very small and insignificant, a characteristic of the Gorgonocephalidœ in contrast to the Trichaster idee. The area just inside the interbrachial ventral surface is occupied by a mosaic of supplementary plates, a characteristic of the Astro- tominœ in contrast to those genera of the Gorgonocephalinœ with arms simple or divided a few times, and a common character of the Astrotominœ and those genera of the Gorgonocephalinœ with arms divided many times. The common character just mentioned may 1)9 due to convergence. PI. II, fig. 11 shows a compound hook of an arm annulus of As',rothamnus echinaceus. Such a compound hook with more than two supplementary booklets besides the terminal one is imcommon among the Gorgonocephalidœ.. The compound hooks of the arm annuli of Astrotoma sobrinn belong partially to this type with two or three supplementary booklets and j)artially to the common type with only a single supplementary booklet besides the terminal one. PI. Ill, figs. 1-6 illustrate the internal structures of the Ophiacanthidœ. The peristomial plates of Ophiacantha hidentata (fig. 1), Ophiolehes tu'jerosa (fig. 4), Ophiolimna antarctica (fig. 5) and 0. papUlata (fig. 6) are almost or perfectly entire and very large, characters common to most of the genera of the Lœ.mophiurida and certain groups of the other orders. That the peristomial I)lates of the two species of Ophiolimm just mentioned are almost entire and very wide and short, is a proof of their Ophiacauthine, and not Ophiodermatine, affinity ; the peristomial plates MONOGRAPH OF JAPANESE OPHIUROIDEA. 387 of the Opldodennntkht; being always triple and rather long relatively to their width. The oral frames of Ophiolimni are very slender, a character of those forms of the Ophincinthidcr with zygospondyline vertebral articnlation ; those of Ophlolehes and Ophiaamtfui. hidentata are rather stout, a character of those forms of the same family with streptospondyline vertebral articula- tion. The vertebral articulation of the latter type is illustrated here by Ophiacantha hidentata (tigs. 2 A' 3). It is noteworthy that both the types of vertebral articulation are represented in Ophwcardha ; Ophiacantha cuspidata having zygospondyline vertebral articulation according to Lyisian. The streptospondyline articulation is a common character of a part of the Ophiaeanthi- dci' and all the Hemîeiiryalidœ among the Lcemophiwlda, as well as of the Ophiohxjrsinœ, Trichastervdai and Gorgonocepftcdkkc among the Fhrynophturida. Ophiolehes tuber osa (fig. i) has short and stout vertebra?, whose wings are rather thick laterally as well as dorsally, also a common character of those forms with streptospondyline vertebral articvilation. The articulation of the genital plate and radial shield is, as shown in the figure, very simple, without any special condyle and socket, a common character of the rhrynop)hmnda, and La'mophiurlda. The genital plate and scale merely articulate with each other, without being soldered together, a character of the Ophiacunlhidœ and some other orders in contrast to the Hemieuryalidcr. PI. ni, figs. 7 & 8 illustrate the internal structures of Amphktctis umbonatu, a representative of the Amphilepklidcr. The iieristomial plates are almost simple and very large, a character distinguishing the present family from all the other GnathophiurhJa. The oral i)lates and frames are slender and the latter have no lateral wungs, also a character distinguishing the present family from the other Gnathophhirkla. The dental plate is absent ; this is a peculiar character. The teeth are not very stout and thick, though appearing wide in tlie figure, also a character uf the present family in contrast to the other Gnathophiurida. The genital plates bear a very conspicuotis, ball-like articular condyle on the dorsal side near the outer end ; this is a charac- teristic of the G'nathophiurkîa in contra-Ht to all the other orders. The genital scales are wide, thin and leaf-like and articulate with the genital plates near the outer ends of the latter ; these characters of the genital scales are notable in the Gnathophiurida, though also found in some ChUophiurkla. The dorsal surface of the vertebra is rhomboidal, instead of being strongly notched at the inner end, a common character of the Amphilepklkid' and Amphiwida; in contrast to the Ophwirichklce among the Gnathophiurida . The distal vertebne (fig. 8) are in this species imperfectly divided into halves by a series of pores. Such vertebrœ or those in which the halves are divided by a moniliform pore, are found in several of the genera with the arms perfectly protected by the plates (see also PI. V, fig. 3) ; while vertebr;e with the halves divided by a fusiform pore are found in the Ophiomyxinœ, in which the dorsal side of the arms is largely unprotected. See text-fig. 2 f & g, p. 18. PI. Ill, figs. 9-13 and PI. IV, figs. 1-5 illustrate the internal structures of the Amphiuridcv. The peristomial plates of Ophiactis pteropoma (PI. ITT, fig. 9), OphiophoHs aculeata (PI. Ill, fig. 10), Hemipholis elongata (PI. IV, fig. 1), Amphioplus ancistrotus (PI. Ill, fig, 11), Amphlacantha diindua (PI. Ill, fig. 12), Ophiophragmus japonicus (PI. IV, fig. 3) and Amphlpholis kochii (PL IV, fig. 2) are simple and very small, a common character of the Amphiurldcc, with the excep- 388 AIIT. 2. — H. MATSUMOTO : tion of the A7nphiiir(i,-{i,toni>, and the Ophiotrichidar, those of Amphiura vadicola (PI. Ill, fig. 13), A. trachydisrji, (PI. IV, fig. 4) nnd Ophiocentrus verticUldtas (PI. IV, fig. 5) are double and very small, a character of the Amphiura-gronp. The common outline of the oral and dental Ijlates of all the representatives here illustrated is X-shaped, thoiigh that of Amphiacnntha dlviduii is slightly atypical owing probably to the fact that the specimen is very small and young. This diaracler is common to the Amph'mridcv and Ophiotrichidc; in contrast to those groups of the ChUophiurida with well developed masticatory apparatxis. The teeth are very sloiit, squarish, with wide and tnuicated tips, and the oral frames are very stout and have well developed lateral wings, charactexs common to the Amph'mrlda;, Ophiotrichidfr, Ophioceramis, Ophumereidinœ and Ophiocomida;. The genital ])lates in Ophiopholis acideato. (fig. 10) and Ophio- centrus verticilldtiis (fig. 5) bear each a very conspicniovis, ball-like articular condyle on the ), 0. koi-enna, (fig. 7) and Ophiothela dana; (fig. 8), the peristomial i)lates are simple and small, a common character of part of the Amphhiridii; and the OpMof.rirhidn; ; the common outline of the oral and dental plates is X-sbapod, a common character of the Amjihliirida; and Ophiotrlchidd' • the teeth are very stout, s<]Uarish, with wide and truncated ends, and the oral frames are very stout and have well developed lati-ral wings, characters common to the Amph'mrldd', Ophintrlcfddti;, Ophioceramis, Ophionereidimr and OjMocoiiudd; ; the genital jilates bear a very conspicuous, ball-like articular condyle on the dorsal side near the outer end, a characteristic of the Gnathophiurldu ; the dorsal surface of the vertebra (except the very basal ones) is Y-shaped, a characteristic of the Ophio- trichidc: in contrast to the .\mphUe2Hdidœ and Amphiuridd; • the genital scales are wide, thin and leaf-like, a character of the (hvdhopk'mridd and part of the Chilophiaridd. The radial shields are, as seen in fig. 6 {Ophlidhrlx wrcidhk'i), very largo and triangular, a common character of most of the genera of the OpJiiotrichidd!, and have a very largo articular socket to fit to the ball-like articular condyle of the genital plate, a characteristic of the (Jii'Uhophiurida. PI. V, figs. 1-5 show the internal structures of A.sirophiiira Jmwumurni. The peristomial jilates are absent, though the peritoneal membranes of the oral region and other parts contain very fine scales (fig. 1). It is stated by L\'M.\n, that Ophiopliidhus medusiv, lacks the peristomial ] liâtes. The oral frames are very long and slender, also reminding us. of those of Opkioplivihus iiifdusd;. The oral plates are al.so slender, ami the dental plate is stout. The genital MONOGUAPil OF JAPANESE OrillUnoiDEA. 389 plates nnd seules are entirely internal and very much reduced in size ; the latter are entirely proximal to the former, the two articulating with each other terminally. These features of the genital plates and scales are (luite peculiar. The genital pLite has no special condyle and socket for articulation with the radial shield, but the outer end itself of the plate serves for an juticular facet. It may bo thought, that Aslrophlnra, is not Chilophiuridan but Tjio- mophiuridau in tlie articulation of the genital plate and radial shit^ld ; Irat my view is that, this condition is simply a result of the extreme reduction of the genital plates. The very l)asal vertebri« are large and long, instead of being very small and short, and are quadrangular in dorsal view, again reminding us of those of OphiopUnthus medusa;. The ventral side of each vertebra within the pentagonal ♦' .-isteroid " body has a well marked medinn suture (fig. 2). The vertebra) of the free arms are dixaded into halves by a very narrow moniliform «lit (fig. 3). The articulation of the same is zygospoudylino (figs. 4 & 5), a (i^mrnon character of the OjMoimJx'mo' , Opkùtcmithidœ pars, thvdhophmriäd and Chilophiurida. VI. V, figs. 6-12 illustrate the internal structures of tlic more typical genera of the Ophiolepidkia: In all the species here illustrated, viz. Ste . . . .Lateral arm plate. liF . . Lower muscular fossa. M. . . . Madreporic oral shield. o . , . . Oral plate. OS . . . Oral shield. p . . , . Peristomial plate. li . . Radial shield. s . . . , Genital scale. u . . . . Articular umbo. UF . . Upper muscular fossa. Y . . , . Ventral arm plate. MONOGRAPH OF JAPANESE OPHIÜKOIDEA. 393 INDEX OF GENERA AND SPECIES. Amphiacantha. . . .150, 168, 177, 178, 353. acanthia 151, 178, 179, 361. amphacantlia 178. dividim. . . . 151, 152, 178, 179, 180, 361, 387, 388. formosa 178, 179, 357. gastracantha 178. libera 178. notacantha 178. pentacantba 1 78. sexradia 178. Amphiactis....4, 91, 145, 146,165, 352. canescens 147. dissidens 147. duplicata 147. X^arata 147. partita 147. patula 147. umboData 146, 147, 361, 387. Araphichilus 150, 168, 175, 353. daleus 175, 177. intermedins 175, 177. triclioides 175, 361. Amphilepis 145, 353. norvegica 146. tennis 146,361. AmiDhilimna 150, 165, 168, 178. pentacantba 178. Amphiocnida 151, 165. Ampliiodia . . 151, 165, 166, 167, 168, 172, 174,182,199, 352, 378. ancistrota 171. craterodometa 182, 355. digitula 167, 199. euryaspis 182, 354. macraspis 173. periercta 184. psilocliora 182, 361. Amphiophinra. . 5, 234, 248, 249, 261, 262,266,353. abcisa 262. abdita ...261. bnllata 261. convexa 261,262,263,356. distincta 261. hadra 262, 266. improva 261, insolita 261 . lacazei 261 . lapidaria 261, 263, 265, 363. latro 261. laudata 261. liberata 261. megapoma 262, 263, 264, 362. obtecta 262. oediplax 261, 262, 263, 362. ornata 262. panpera 261. penichra . . 262, 263, 264, 266, 362. pompophora . . 261, 262, 264, 362. ponderosa. .261, 263,264,266,359. prisca 261. radiata 262. remota 261. sculptilis 261, 263, 265, 356. scutata 261. solida 249,261. sordita 261. stellata 261. 394 ART. 2. — H. MATSÜMOTO undata 261. urbaiia 261. Ampbioplus . . 147, 150, 165, 166, 167, 168, 169, 172, 174, 175, 178, 200, 353. acautliiuus 178, 179. ancistrotus 151, 152, 170, 171, 361, 387. cernuus 170, 172, 361. glaucus 170, 173, 361. hexacauthus 170, 172, 358. macraspis 170, 173, 358, megapomus 169, 170, 361. rbadiuobracliius .... 169, 171, 358. Amplnpholis . 151, 165, 166, 167, 168, 175, 198, 208, 352. audrese 184. australiana 188. corese 198. japonica 185, 186, 190, 361. kocliii .... 151, 185, 192, 361, 387. lütkeni 208. pugetana 185, 191, 358, 365. sobrina 185, 189, 362. squamata 187, 188, 1 90. Amphipholizona 235. Ampliipsila 93, 375. Ampbiura . . . 147, 151, 152, 165, 166, 167, 168, 169, 174, 194, 198, 199, 352, 378, 388. acacia 196, 203, 362. acrystata 194, 196, 203, 359. sestuarii 197, 208, 362. anoniala 174. assimilis 174. bellis -. . 195, 200, 357. caudita 201. cauesceus 147. carcbara 196, 203, 358, 365. cernua 172. corese 198. crassa 174. dalea 175. digitula 174, 195, 199, 362. diomedese 198, 199. duncani 208. duplicata 147. ccnomiotata 198, 210, 213, 362. euopla 196, 201, 210, 362. gastracantba 178. glauca 173. grabra 205, 207. grata 174. intermedia 175. iridoides 197, 205, 362. iris 197,204,357. kocbii 192. koreœ . . 151, 152, 174, 195, 198, 356. lütkeni 197, 208, 357. micraspis 197, 208, 362. microdiscus 195, 201, 362. uotacantha 178. partita 147. patula 147. pbarelata 210, 211, 213. prsestans 184. pugetana 191. pycnostoma 196, 203, 362. reposita 174. servata 174. tracbydisca 151, 152, 195, 201, 362, 388. vadicola. . . . 152, 197, 210, 211, 362, 388. Autbopbiura. .234, 249, 250, 382, 383. Aspidopbiura .... 234, 248, 249, 250, 353, 382, 383. MONOGKAPH OF JAPANESE OPfllUROIDEA. 395 forbesi . . 249, 250, 252, 253, 357. mimita 249,252. watasei 249, 250, 362. Asterias — acnleata 161. bideutata 118. opbiura 267. Asteronyx 4,28,353,373. loveni 29, 33, 355, 385. Asteroporpa 4, 59, 67, 353. badracantba .... 61, 62, 63, 67, 359. Asteroscbema 4, 29, 42, 43, 44, 49, 54, 353. abyssicola 42, 53, 359. caudatum .... 33, 42, 49, 52, 53, 55, 359. enosbimanum 51, 52. glaucum 42, 46, 55, 359. glutinosum 42, 53, 55, 359. bemigymnum 42, 47, 55, 359. iutestum 49. japonicum . . 32, 42, 51, 53, 55, 359, 386. kœbleri 29. migrator 49. monacantbum 51, 52, 53. œdipus 49, 51« rubrum 45. sagammum 49, 50. Salix 47. tubiferum 42, 44, 55, 359. Astroboa 60, 67, 79, 353. arctos 61, 65, 79, 80, 360. clavata 83. evûse 83. globifera 79, 83, 84, 360. nigra 79,84,357. nuda 83. Astrocanenm 60. Astroceras 4, 28, 29, 30, 32, 35, 54, 55, 353, 386. adbereus 29. kœbleri 29. pergamena 30, 35, 55, 356. Astrocbalcis 60. Astrocbaris 29, 42, 353. ijimai 56, 359, virgo 57, 58. Astrocbele 4, 59. lymani 61, 62. Astrocblamys 59. Astrocladus 4, 60, 67, 74, 353. anunlatus 61, 64, 74, 75, 360. coniferus 61, 65, 74, 77, 357. dofleini 74, 77. pardaHs 74, 77. Astroclon 61, 84, 353. snensoni 90, 360. Astrocnida 59. isidis 61, 62. Astroconiis 60. Astrocyckis 60. Astrodactybis 60. Astrodeudrum 4, 60, 67, 68, 353. sagaminnm 61, 64, 73, 357. Astrodia 28, 34. Astrogeron 8, 91. Astrogompbns 59. vallatus 62. Astrogordius 60. As'rogymuotes 29. Astropbiura .... 4, 91, 234, 236, 237, 239, 353, 378, 382, 389. cavelljB 239, 245, 246. kawanmrai . . . .236, 239, 362, 388. permira 245, 246. Astropbytou 60. 396 ART. 2. — H. MATSUMOTO : asperum 40. caryi 71. clavatum 79. coniferum 77. coruutiim 77. globifernm 84. pardalis 77. stimpsoni 72. Astrophytum 60. Astrothamnus 5, 60, 84-, 353. bellator 85, 86. ecliinacens .... 61, 65, 85, 86, 360, 386. rigens 85, 86. vecors 85, 86. Astrothoras .... 61, 84, 353. misakiensis 67, 87, 360. Astrotlirombus CO, 84. Astrotoma 4, 61, 84, 353. agassizi 67. murrayi 88, sobriua 61, 66, 88, 360, 386. Astrospai'tns 60, 67. Batliypectinura . . 309, 319, 333, 353. gotoi 309, 320, 326, 363, 390. lacertosa 321. reducta 333. Conocladns 59. Cryptopelta 308, 309, 311, 312. Ctenamphiura 151, 169. Diopederma 309. Euryala — anopla 42. aspera 41. Enryale .... 4, 29, 30, 32, 35, 40, 353. anopla 40, 41, 359. aspera .... 30, 31, 40, 61, 356, 366. aspernra 40. parmiferum 37. stiideri 40. Gorgonoceplialus .... 4, 60, 63, 67, 69, 70, 352, 373. agassizi 61, 68. arcticns 61, 64, 68. caryi 61,64,70,71,354. cliileusis 61, 64. dolicLodactylns . . 61, 64, 70, 73, 359, 865. japonicns 72, 73. sagaminns 72, 73. stimpsoni 72. tnberosns .... 61, 64, 70, 359, 365. Gymnolopluis 216. Gymnopliiura 234, 265. cgelurescens 266, 273, 274. mollis 266. Haploplnura 5, 234, 246, 293. gymnopora 237, 247. Hemienryale 70, 139. pnstulata 140. Hemilepis 163, 194. Hemipholis 150, 165, 201, 378. elongata 151, 152, 387. microdiscus 201 . Homalopliiura .... 234, 248, 249, 266, 352, abyssoriim 266. brncei 267. clasta 266, 363. confragosa 266. divisa 267. flexibilis 267. frigida 267. gelida 267. inermis 267, 293. inflata 267. inortata 266. intorta 266. MONOGRAPH OF JAPANESE OPHIÜR0IDEA. 397 mimaria 267. nana 267. partita 267. rouclii 267. scissa 267. scntellata 267. tessellata 266. Lütkeuia 216. Microplnura . . 4, 91, 93, 95, 375, 382. Neoplax 9, 14, Opbiacantha 94, 96, 98, 106, 107, 111, 317, 352, 387. acanthinotata 112, 117, 360. adiaphora 112, 117, 355. aenigmatica Ill, 113, 360. auchilabra 112, 117, 360. armata 116. bairdi 101. bidentata. . . .96, 112, 118,354, 386, 387. bisquaraata 112, 120, 360. cataleimmoida 106, 108. cuspidata 96, 387. dallasii Ill, 113, 357. diploa 112, 119, 360. euphylactea 110. eurypoma 106. gracilis 125. grönlandica 118. hylacantha 106, 109. inutilis 112, 117, 357. lambda 101. leucorliabdota 101, 109. leucosticta 110, 111. levispina Ill, 115, 357. lineolata 317. lopliobracliia 112, 117, 360. megatrema 100. normani 106, 109. œdidisca 136, 137. omoplata Ill, 116, 360. pentagona Ill, 116, 355. perfida 101. prionota 112, 119, 360. relicta 101. rliacbopliora 112, 119, 355. rosea .112, 122,355. scolopendrica 110, 111. spinulosa 118. trachybactra 115. Ophiacanthella 93, 98, 353. acontophora 123, 355. troscheli 123. Opliiactis . .4, 147, 150, 153, 353, 378. affinis 153, 155, 357. brachygenys 153, 154, 361. clavigera 136. dissidens 147. dyscrita 158, 154, 361. gymnochora .♦. 154, 159, 361. krebsii 158. raacrolepidota 153, 155, 357. modes ta 153, 156, 357, 366. parata 147. profundi 154. pteroporaa .... 151, 152, 153, 154, 361, 387. reinliardti 158. savignyi 154, 158, 355. sexradia 158. Ophialcsea 94, 107. Ophiambix 93, 97, 353. aculeatus 179, 357. Opliiarachna . . 308, 311, 317, 319, 330, 353. danbyi 319. gorgonia 323. incrassata. . 309, 316, 318, 326, 358. 398 ART. 2. — H. MATSUMOTO 366, 376, 390. inferualis 324. Opliiaraclmella. . . .309, 319, 322, 353. gorgonia . .303, 309, 322, 323, 358, 366, 390. infernalis 323, 324, 358, 366. megalaspis 323, 325, 363. Oi3lnartLram 341, 343, 353. elegaus .... 342, 351, 358, 366, 391. Ophientrema 93, 98,110, 353. eupliylactenm 110, 360 lencostictnm 110. scolopeudricum 110, 111, 360. Ophiernus 304, 305, 353. adspersus 307, 356. aunectens 307. vallincola 305. Ophiosethiops 216, Ophiobleiiua 93, 375. Ophiobrachion 9, 25, 373. Oijlnobyrsa 9, 23, 352. acantliinobracliia 24, 359. riidis 11. syuaptacantlia î^4, 359. Ophiobyrsella 9. Opliiocaraax .... 94, 97, 99, 131 , 353. Lystrix 96, litbosora 131. polyploca 132, 361. Opliiocampsis 215. Ophiocentrus .... 151, 152, 165, 169, 353. verticillatns 152, 213, 357, 366, 388. Ophioceramis . . 4, 236, 238, 381, 382. 388, 389, 390, 391. ' januarii 238, 389. obstricta 140. Ophiochseta 5, 309, 317- mixta 101, 316. Ophiocliasma 309. Opliiochitou . . 297, 326, 328, 336, 353, 380, 381. cariuatus 329, 330. fastigatus . 316, 326, 327, 328, 356, 390. lymani 326. Ophiocliondrella .... 93, 97, 136, 375, 376. squamosa 97. Opluochondrus 139, 140. armatus 139, couvolutus 139. gramilatus 139. Opliiochrysis 234. Ophiochytra 145, 147. Ophiocimbium 93, 375. Ophiocirce 304. Ophiocnemis 215, 217. Opliiocnida 151, 165, 168. amphacantha 178. libera 178, liitkeni 208. sexradia 178. verticiUata 213. Ophiocoma. . . 341, 342, 343, 349, 353, 381. arctica 118. brevipes . . 342, 343, 345, 358, 366. echinulata 1 18. erinaceus 343, 315, 346, 347, schceDleinii . . . .343, 345, 346, 347, scolopendrina . , 342, 343, 345, 346, 347, 358, 366, 391. Ophiocoûis 4, 308, 310, 311. antarctica 101. bvevispiua 310. ciucta 310, 315. MONOGRAPH OF JAPANESE OPHIUBOIDEA. 399 cupida 310, 315. diastata 101, 310. forbesii 310. graudisqnama 310, 315. indica 310,311,312. miliaria 310, 312. papillata 101, 310. permixta 310, 315. pulverulenta 310, 312. Ophiocopa 93. Ophiocormus 308. Ophiocrasis . . 326, 327, 328, 334, 337, 353, 390, 391. dictydisca 337, 340, 363. marktauueri . . . .327, 334,338, 363, 390. Ophiocrates 235. Ophiocreas 29, 43, 49. abyssicola 53. adherens 29. caudatus 49. japonicus 51. œdipns 49, 51. papillatus 5 J , 52. Ophiocrene 60, 68. Ophiocryptus , . 309. Opliiocten 235. brevispinum 306. charischema 306. hastatum 326. lymani 326. oöplax 268, 275. pacificum 326. sericeum 237. Ophiocynodus 8, 369. Ophiodera 9, 13, 352, 369. anisacantha . . 10, 21, 359, 371, 384. Ophioderma 298, 309. brevicauda 309. jannarii 309. Ophiodoris . . 297, 326, 327, 333, 353, 380, 390, 391. pericalles 328, 334, 363, 390. Ophiogeron 3, 8, 91. Ophioglypha 247. abcisa 262. abdita 261. abyssorum 266. acervata 267. îeqnalis 267- albata ....268,281. ambigna 268. ampbitrites 267. aspera 2fj8. aurantiaca 267. brevispina 267. brucei 267. bnllata 261. carinata 282. cleraens 268. concreta 268. confragosa 266. convexa 261, 263. costata 268.. deshayesi 238, 282: distincta 261. divisa . , 267. elevata 254. falcifera 268, ferruginea 271. flageUata 267,273. flexibilis 267. forbesii.... 249, 253. fraterna 268. frigida 267. gelida 267. grandis 277. hexactis 235. 400 ART. 2. — H. MATSUMOTO imbecillis 267, 276. improva 261. iudica 267. inermis 267. inflata 267. inortata 266. insolita 261. intorta 266. involuta 277. irrorata 268,277. jejuna 268. kinbergi 267,271. lacazei 261. lapidaria 261, 265. latro 261. laudata 261. lepida 267. liberata 261. Ijungmani 267. loveni 268. lütkeni 267. Ijraani 268. maculata 267, 276. meridionalis 267. mimaria . . . , 267. minuta 249. multispina 268. mundata 268. nana 267. nodosa 258. obtecta 262. orbiculata 277. ornata 262. papillata 267. partita 267. paupera 261. plana 268. ponderosa 261, 264. prisca 261. radiata 262. remota 261. rouclii 267. rugosa . , 268. sarsii 272, scissa 267. sculpta 254, 258. sculptilis 261, 265. scutata 261. scutellata 267. sinensis 271. sladeni 254, 259, 260. solida 261. sordita 26 1 . stellata :]61. sterea 254, 258. striata 254, 255. tenera 268. tessellata 266. thouleti 267. tumulosa 277. undata 261. undulata 268. urbana 261. variabilis 265. verrucosa 268. Opliioglyphina 247. Ophiogona 235. Opliiogymna 216, 217. Ophiogyptis 139. nodosa 143. Ophiohelus 3, 8, 91. Ophioholcus 139. Ophiohyalus .... 3, 9, 12, 16, 91, 353, 369, 371. gotoi 11, 17, 359, 384. Opliiohymen 9, 13, 353, 369, gymnodiscus 23, 359. Ophiolebes . .93, 97, 99, 115, 132, 136, MONOGKAPH OF JAPANESE OPHIUROIDEA. 401 352, 375, 376, 387. jirmata 139. asaplies 132, 135, 361, 365. brachygnatha . .132, 135, 361, 365. brevispina 136. clavigera 136. diaphora 136. granulata 139. pachybactra 136. paucispina 136. tuberosa .... 96, 133, 361, 365, 386, 387. tylota 136, 138. Ophioleda 94, 125, 126. miuima 124, 126. Ophiolepis . . 235, 282, 283, 292, 298, 298, 299, 353. annulosa 299, 300, 358, 366. carinata 282. cincta 238, 299, 358. gaiTettii 299. imbricata 301. impressa 298. mirabilis 160. nigra 299. pacifica .... 298. robusia 267. savignyi 158. sexradia 158. superba 300. Opliioleptoplax . .9, 12, 352, 369, 371. megapoma 13, 359. Ophioleuce 304, 305, 306, 353. brevispinum 306, 363. chariscliema .... 304, 306, 363, 389. Opliiolimna . . 5, 93, 97, 101, 102, 310, 311, 316, 353, 375, 387. autarctica 95, 101, 386. bairdi 101, 102, 355. diastata 101, 102, 310, 360. lambda 101, 102, 103, 360. littoralis 101, 316. mixta 101, 316. operculata 101. papillata 95, 101, 310, 386. perfida 101. Ophiolipus 235. Opliiologimus . . 4, 91, 92, 95, 97, 353, 369, 375. hexactis 99, 360. Ophiomastix 341, 343, 348, 353. armulosa . . 342, 348, 350, 358, 366, 391. lütkeui .... 342, 348, 349, 358, 366. mixta .342, 348, 358, 366. Oplnomastus . . 234, 236, 237, 378, 382. secundus 237. Ophiomaza 215, 217, 353. cacaotica 230. kanekoi 227, 362. mcerens 230. obscura 230. Ophiomedea 93. Ophiomidas 235. Ophiomisidimn .... 234, 237, 378, 382. Ophiomitra 94, 99, 131, 353. acontophora 123. bytliiaspis 131, 361. cardiomorplia 130. codonomorpha 106, 109. dicycla 130. exigua : 123, 130. granifera 106. habrotata 124, 126. lithosora 131, 361. micropliylax 106, 110. normani 106, 109. 402 ART. 2. — H. MATSUMOTO polyacantha 189. Ophiomitrella . . 93, 98, 103, 109, 353. americana 106. languida 106. mutata 106. partita 105. j)lacicla 106, polyacantha 103, 106, 360. stellifera 103, 360. Ophiomœris 139, 140, 353. obstricta 140, 143, 357. projecta 139, 140, 141, 361. Opliiomora 9. Ophiomusium .... 235, 238, 283, 284, 353, 378, 382, 383, 389. caucellatum .... 238, 285, 291, 356, 389. granosum 284, 289, 363. jolliense 285, 291, 359. laqiieatum 285, 291, 358. lunare 284, 289, 358. liitkeni 285, 290, 358. lymani 284, 289, 356. sanctum 288. scalare 284, 285, 356. Simplex 234, 288, 358. trychnum . . 238, 285, 290, 363, 389. Ophiomyces .... 92, 97, 353, 3G9, 375. frntectosus 96. spathifer 99, 360. Opbiomytis 94. Ophiomyxa. . 3, 9, 13, 16, 91, 352, 369, 371. anisacantha 21. australis 10, 19, 356, 384, flaccida 10, 376. Ophioncns 309. Ophionema 151, 165, 169. Ophionephthys . . . 151, 165, 194, 203. phalerata 211. Ophionereis. .326, 327, 328, 333, 334, 337, 353, 390, 391. annulata 327, 390. dubia 334, 336, 337, 340. eurybrachiplax 334, 336, 359. porrecta 334, 338, 340, 358. reticulata 327, 390. sinensis 334, 336, 363. squamata 335. variegata 335. Ophionotus 235, 268. hexactis 235. Ophiopsepale 304. goësiana 305. Ophiopallas 304. Opliiopeltis 165, 194. Opliiopenia 235, 283, 352. disacantha 292, 355. Ophioperla 235, 268. Ophiopeza — danbyi 319. reducta 333. Ophiopezella 309. Ophiopholis ... .3, 150, 153, 159, 352, 372. aculeata . . .151, 152, 159, 161, 162, 354, 387, 388. brachyactis .... 160, 163, 361, 365. japonica ... 160, 162, 163, 164, 354. mirabilis 159, 160, 361, 365. Ophiopbragmus . . . 151, 165, 168, 353. affinis 184. andrese 184. japonicus . . 151, 152, 183, 361, 387. perierctus 184, 185. preestans 184. Ophiophrixus 9, 23, 25, 353. MONOGRAPH OF JAPANESE OPHIUROIDEA. 403 acauthinus 24, 359. Ophioplirnra 93, 97, 353. liodisca 99, 360. Ophiophtlialmns ... .93, 98, 106, 107, 137, 353. americanus 106. cataleimmoidas 106, 107, 108, 355. codonomorplia . . 106, 107, 109, 360. eurypoma 106. graniferus 106. hykcantha .... 106. 107, 109, 360. langnidiis , . 106. leucorhabdotns . . 106, 107, 109, 360. micropliylax .... 106, 107, 110, 360. mutatus 106. normarii 106, 107, 109, 354. placidus 106. relictus 106. OpLiophthyreus 215. Ophiophycis 234, 237, 382. OplnopLyllum 235. Ophioplax .... 308, 326, 328, 353, 381. custos 333. lamellosa . . 316, 327, 330, 363, 390. Ijiingmani 333. reducta 333. Opliiopleura 235. Ophioplinthaca 94, 99, 353. cardiomorpha 130, 361. occlus 1 124. Ophiopliuthus 235, 236, 237. medusa 237, 388, 389. Ophioplocus. .235, 284, 301, 353, 389. imbricatiis 301, 302, 303, 358, 366. japonicus . . 238, 301, 302, 363, 389. tessellatus 301. Ophiopius 139. armatus 139. Ophiopora 93, 97, 100, 353. bartletti 100. megatrema 100, 360. paucispina 100. Ophiopristis 93. Opbioprium 93. Ophiopsammium 216, 217. Opbiopsila .... 4, 341, 342, 378, 391. riisei 342, 391. Ophiopteris 341. Ophiopterön 215. Opbiopus 150, 378. arcticus 95, 152. Opliiopyren 304. lougispiuns 305. Oplnopyrgus 234, 262, 382, 383. Opliioscalus 93. Ophioschiza 9. Ophiosciasma 3, 8, 91, 369. Opliioscolex 9, 14, 62, 372. glacialis 11. Ophiosemnotes 5, 93, 97, 99, 135, 136, 352, 375, 376. brevispina 136. claviger 136. diaphora 136, 138. œdidisca. ... 96, 136, 137, 361, 365. pachybactra 136. paucispina 136. tylota 136, 137, 138, 355. Ophiosmilax 9, 24, 353. mirabilis .... 11, 25, 359, 384, 385. Ophiosphœra 216. Ophiosteira 234. Ophiostiba . . 3, 9, 12, 13, 91, 353, 369. hidekii 11,14,359,384. Ophiostigma 151, 165, 168, 179. formosa 178, 179. 404 ART. 2. — H. MATSUMOTO : Ophiostyracium 8, 12, 353, 369. tracliyacanthum 13, 359. Opliiosyzygus 8, 12, 353, 369. disacanthus 13, 359. Ophioteresis 216. Opliiotliamnus . . . .4, 91, 94, 99, 123, 124, 125, 353. affinis 126. exiguns 123, 130. gracilis 125. habrotatns 124, 126, 360. Isevis 123, 130. minimus 124. remoüls 126. stultus 123, 130. vemistus 94, 126, 361. vicarius 94, 126. Ophiothela 216, 217, 353. danœ 217,230,232,357. involuta 231. isidicola 231,232. verriUi 231,232. Opliiotholia 92. Ophiotbrix . . 212, 215, 217, 218, 353. catapliracta 224. cheneyi 226. ciliaris 224. crassispina 222, eusteira 218, 222, 362. galapagensis 223. liirsuta . . . .219, 225, 226, 357, 366. hylodes 221. koreana . . 155, 217, 218, 220, 222, 357. longipeda 219, 227, 357, 366. macrobi-acliia 218, 222, 362. marenzelleri 218, 220, 362. nereidina 216, 217, 218, 224, 357, 366. obtusa 219, 225, 357, 366. paDchyendyta 218, 219, 362. punctolimbata 219, 226, 357. stabilis 218, 224, 362. stelligera 220. variabilis 226. Opliiotliyreus 235. Opliiotjalfa 235, 268. Opliiotoma 93, 375. Ophiotrema 93, 100. alberti 100. Ophiotreta 317, 318. Opliiotrichoides 215. Ophiotroclms .... 304, 305, 353, 382, 383. longispimis 307, 363. Opbiotypa 234, 237, 382. Oplnozoua . . 148, 235, 282, 292, 293, 298, 336, 381, 389. alba 294. antillarum 294. bispinosa 294. capeusis 268, 293. casta 294. clypeata 294. contigua 294, depressa 294. elevata 294, 295. gymnopora 247, 292. impressa 238, 298, 389. iuermis 267, 293. iusularia 294. loDgispina 294, 297. marmorea 294. media 294. molesta 294. uivea 294. pacifica 298. platydisca 291,295. MONOGRATH OF JAPANESE OPHIUROIDEA. 40^ polyplax 294,290. projecta 294,295. steUata 294. tesseUata 294. tjalfiana 294. Oplnozonella . . 5, 235, 238, 283, 293, 294,336,353,38], 389. alba 294,297. antillarum 294. bispinosa 294. casta 294. clypeata 294. contigua 294. depressa 294. eleyata 294, 295, 363. insnlaria 294. loDgispina .... 238, 294, 295, 297, 363, 389. marmorea 294. media 294. molesta 294. nivea 294 . platjdisca 294, 295, 363. polyplax 294, 295, 296, 363. projecta 294, 295, 358. stellata 294. tessellata 294. tjalfiana 294. Ophiozonoida 235. Ophiura. . 235, 247, 248, 249, 266, 268, 275, 282, 305, 852, 382, 389. acervata 267. œqualis 267. affinis 267. albata 268, 270, 281, 363. albida 267. ambigua 268. amphitiites 267. anniilosa 300, 350. arctica 267. aspera 268. atacta 268. aurantiaca 267. bathybia 268. bracliyactis 254, 260. brevispina 267. calyptolepis 268, 270, 281, 363. capensis 268, 293. carnea 267. clasta 266,267. clemens 268. concreta 268. costata 268. cryptolepsis 268, 270, 281, 354. falcifera 268. flagellata . .266, 267, 269, 273, 356. fraterna 268. glyptodisca 249, 253. liadra 262. imbeciUis 267, 269, 276, 358. incrassata 318. indica 267. inermis 267. irrorata . .249, 268, 270, 277, 280, 356. jejuna 268. kinbergi . . 237, 238, 267, 269, 271, 357. lepida 267. leptocteuia .... 268, 270, 277, 354. Ijnngmani 267. longipeda 227. loveni 268. Kitkeni 267. lymani 238, 268. maculata 267, 269, 276, 355. 406 ART. 2. — n. MATSUMOTO raegapoma 262, 264. meridionalis 267. micracantha .... 268, 269, 276, 363. monostoecha .... 268, 270, 281, 363. raultispina 268. muudata 268,280. uereidina 224. nodosa 254, 258. œdiplax 261, 263. oöplax 268, 269, 275, 363. ophiura 267, 268. papillata 267. paucisquama . . 268, 270, 278, 363. penichra 262, 264. plana 268, 280. pompophora 261, 264. ponderosa 264. quadrispina .... 268, 269, 275, 355. retzii 118. robusta 267. rugosa 268. sarsii 267, 269, 272, 354. scolopendrina 345. sculpta , 258. scnlptilis 265. sterea 258. stipliia 254,259,260. stuwitzii 254. tenera 268. thouleti 267. iindiilata 268. verrucosa 268. Ophiurases 139, 140. obstricfcus 140. Opliiuriua 371. OpLiurocliceta . . 5, 308, 311, 316, 380, 390. littoralis 101, 316. mixta 101, 309, 316, 390. Ophiuroconis 4, 5, 308, 310, 311, 312, 353, 380, 390. miliaria 312, 314, 390. monolepis 309, 312, 316, 363. pulverulenta 312, 314. Opliiurodon ... .4, 308, 310, 311, 314, 353, 380, 390. cinctti 315. cupida 315. grandisquama . . 309, 315, 316, 358, 390. permixta 315. Opliiurolepis .... 235, 249, 281, 282. carinata 238,282. Ophiuropsis 28. Ophiurothamuus .... 94, 99, 125, 129, 353. dicyclus 130,361. exiguus 130. Igevis 130. stultus 130. Paramphiura 151, 165, 169. Pectinura .... 309, 311, 312, 319, 353. anchista 309, 322, 363. cylindrica 322. danbyi 319. gorgonia 323. infernalis 324. intermedia 323. marinorata 323. ramsayi 323. similis 324. stearnsii 323. venusta 323. Sigsbeia 139. Stegophiura. .234, 237, 248, 253, 254, 352,389. brachyactis 254, 255, 260, 362, 365. MONOGRAPH OF JAPANESE OPHIUROmEA. 407 elevata 254. nodosa 254, 255, 258, 354. sculpta 254, 258, 362, 365. sladeni . . . .237, 254, 255, 259, 362, 365, 389. sterea .... 237, 254, 255, 258, 362, 365. striata 254, 255, 362, 365. stuwitzii 254. vivipara 254, 255, 362, 365. Sthenocephalus 28. Trichaster. ... 4, 28, 30, 32, 35, 37, 54, 55, 353, 386. elegans. ... 30, 37, 38, 55, 356, 366, 385. flagellifer 37. palmiferas' 37, 38, 39, 356. 408 ART. 2. — H. MATSUMOTO. CONTENTS. PAGE Introductory 1. I. Œgox^biuroida 5. II. Myopliiuroida 6. i. Pluynophiitrida . . . . . . . . 6. Ophiomyxidse 7. Trichasteridse 27. GorgonocephalidEB • 58. ii. Lremophiurida 90. Ophiacanthidse 92. Hemieuryalidœ 138. iii. Gnatliophiurida 143. Amphilepididse 144. Amphinridse 149. Opliiotricliidœ 214. iv. Chilopliiurida 232. Ox^biolepididsß 233. Ophioleucidse 303. Opliiodermatidœ 307. Ophiocliitonidse 325. Opliiocoinida3 340. OpLiuran Fauna of Japan 352. Phylogeny of Opliinroidea 367. Postscript 384. General Explanation of Plates 384. Index of Genera and Species 393. Plates I — ^VII, with Explanations. Published March 31st, 1917. II. IMATSUMOTO. MONOGRAPH OF JAPANESE OPHIUROIDEA. PLATE I. PLATE I. Fig. 1. OpModera anisacantJia. Dorsal view of the skeleton of two oral augles and one arm base, with the radial shields turned up, i.e. in ventral view. X 6. Fig. 2. Ditto. Inner articular surface of a vertebra near disk, x 6. Fig. 3. Ditto. Outer articular surface of the same. X 6. Fig. 4. Opldomyxa ausiralis. Dorsal view of the skeleton of two oral angles and one arm base of a young specimen, with the radial shields turned up, i.e. in ventral view. X 30; Fig. 5. Ditto. Inner articular surface of a vertebra near disk, x 30. Fig. 6. Ditto. Outer articular surface of the same, x 30. Fig. 7. Ditto. Dorsal view of the skeleton of two oral augles. x 15. Fig. 8. OpMoliyalus gotoi. Dorsal view of the skeleton of two oral angles and the first vertebra, x 15, Fig. 9. Ophiostiba JddeJcii. Dorsal view of the skeleton of two oral angles. Xl5. Fig. 10. OpMosmilax mirahilis. Inner articular surface of a vertebra near disk. X 40. Fig. 11. Ditto. Outer articular surface of the same. x40. Fig. 12. Ditto. A long arm spine, with two series of booklets, x 40. Fig. 13. Ditto. A short arm spine, with one main and one supplementary booklet. X 40. Fig. 14. Asteronyx loveni. Dorsal view of the skeleton of two oral angles and one arm base of a young specimen, with the radial shields turned up, i.e. in ventral view, x 20. Fig. 15. Ditto. Inner articular surface of a vertebra near disk, x 20. Fig. 16. Ditto. Outer articular surface of the same. x20. Jour. Sei. Coll. Vol. XXKVIII., Art. 2, PI. I. Matsumoto del. Matsumoto : Japanese Ophiuroidea. II. MATSUMOTO. MONOGRAPH OF JAPANESE OPHIUROIDEA. PLATE IL PLATE n. Fig. 1. OpJdosmilax mirabilis. Dorsal view of a vertebra near disk, x 40. Fig. 2. Asteroschema japoniciim. Dorsal view of the skeleton of two oral angles and one ann base, with the radial shields turned up, i.e. in ventral view. X 3. Fig. 3. Ditto. Ventral view of the skeleton of two oral angles and one arm base, x 3. Fig. 4. Ditto. Side view of the skeleton of one oral angle and one arm base. X 3. Ditto. Outer view of an arm joint near disk. x3. Ditto. Inner articular surface of a vertebra near disk, x 3. Trichaster elegans. Ventral view of the skeleton of two oral angles and one arm base, x 2. Ditto. luterbrachial lateral view of the skeleton of the disk and two arm bases, x 2. Astrotoma sobrina. Dorsal view of the skeleton of two oral angles and one arm base, especially showing the muscles between the basal vertebrae and the genital plate, x 3. Fig. 10. Ditto. Ventral view of the skeleton of two oral angles and one arm base, x 3. Fig. 11. Astrothamnus echinaceus. A compound hook of an arm annulus. X 35. Fig. 5. Fig. 6. Fig. 7. Fig. 8. Fig. 9. Jour. sa. Coll. Vol. XXXVIII., Art. 2, PI. Il, m 'ö A II Si Matsumoto del. Matsumoto : Japanese Ophiuroidea. Il, MATSUMOTO. MONOGRAPH OF JAPANESE OPHIÜROIDEA. PLATE III. PLATE III. Fig. 1. OiMacantha hidentata. Dorsal view of the skeleton of two oral angles and the first vertebra, x 10. Fig. 2. Ditto. Inner articular surface of a vertebra near disk, x 10. Fig. 3. Ditto, Outer articular surface of the same, x 10. Fig. 4. 02M6lébes tuherosa. Dorsal view of the skeleton of two oral angles and one arm base, with the radial shields turned up, i.e. in vental view, x 10. Fig. 5, OpMolimna antardica. Dorsal view of two oral angles, x 10. Fig. 6. OpMolimna papillata. Dorsal view of two oral angles, x 12. Fig. 7. Ampliiadis umhonata. Dorsal view of the skeleton of two oral angles and one arm base, x 20. Fig. 8. Ditto. Ventral view of a distal vertebra. x80. Fig. 9. OjMadis pleropoma. Dorsal view of the skeleton of two oral angles, x 15. Fig. 10. Ophiopholis aculeata. Dorsal view of the skeleton of two oral angles and one arm base, x 6. Fig. 11. Amphioplus ancistrotvs. Dorsal view of the skeleton of two oral angles, x 18. Fig. 12. AmpMacantha dividua. Dorsal view of the skeleton of two oral angles of a young specimen, x 35. Fig. 13. Ampliiura vadicola. Dorsal view of the skeleton of two oral angles, x 10. Jour. Sei. Coll. Vol. XXXVIII., Art. 2, PI. ///. Matsumoto del. Matsumoto : Japanese Ophiuroidfca ir. 3IATSUM()T(). MONOGRAPH OF JAPANESE OPHIUROIDEA. PLATE IV. PLATE IV. Fig. 1. Hemipliolis élongato^ Dorsal view of the skeleton of two oral angles and the first two vertebrae, x 16. rig. 2. AniiihipJiolis kochii. Dorsal view of the skeleton of two oral angles, x 18. Fig. 3. OpMophragmus japonicus. Dorsal view of the skeleton of two oral angles, x 20. Fig. 4. AmpJdura trachydisca. Dorsal view of the skeleton of two oral angles, x 8. Fig. 5. Ophiocentrus verticillatus. Dorsal view of the skeleton of two oral angles and one arm base. xl3. Fig. 6. OpMothrix nereidina. Dorsal view of the skeleton of two oral angles and one arm base, with the radial shields turned up, i.e. in ventral view. X 10. Fig. 7. Opldolhrix horeana. Doi'sal view of the skeleton of two oral angles and one arm base, x 10. Fig. 8. OpJdotlida dance. Dorsal view of the skeleton of two oral angles and one arm base, x 10. Jour. Sei, Colt. Vol. XXXVIII., Art, 2, PI. IV. Matsümoto del. Matsumoto : Japanese Ophiuroidea. II. MATSIT3I()T(). MONOGRAPH OF JAPANESE OPHIUROIDEA. PLATE V. PLATE V. Fig. 1. AstropJiiura kaioamurai. Dorsal view of the skeleton of two oral angles and one arm base, x 10. Fig. 2. Ditto. Ventral view of the skeleton of three arm joints within the pentagonal body, the ventral arm plates removed to expose the vertebrae, x 10. Fig. 3. Ditto. Dorsal view of a distal vertebra, x 125. Fig. 4. Ditto. Inner articular surface of the same, x 125. Fig. 5. Ditto. Outer articular surface of the same, x 125. Fig. 6. StegopMura sladeni. Dorsal view of the skeleton of two oral angles and the first two veitebrae. x 18. Fig. 7. OpMommium trydmum. Dorsal view of the skeleton of two oral angles and the first two vertebrae, x 10. Fig. 8. Ophiomnsium cancellatum. Dorsal view of the skeleton of two oral angles and the first two vertebrae, x 12. Ophiozonella longispina. Dorsal view of the skeleton of two oral angles and the first two vertebrae. > 6. Ophiozona impressa. Dorsal view of the skeleton of t^o oral angles and the first three vertebae. x 8. Ophioplocus joponicus. Dorsal view of the skeleton of two oral angles and one arm base, x 8. Ophioceramis januarii. Dorsal view of the skeleton of two oral angles and the first vertebra, x 10. Fig. 13. Opldoleuce diariscliema. Dorsal view of the skeleton of two oral angles and one arm base. X 8. Fig. 9. Fig. 10. Fig. 11. Fig. 12. Jour. Sei. Coll. Vol. UXVIII., Art. 2, PI. V. Matsumoto del. Matsumoto: Japanese Ophiir.oidea. II. MATSUMOTO. MONOGRAPH OF JAPANESE OPHIUROIDEA. PLATE VI. PLATE VI. Fig. 1. Ophiurodon grandisqunmn . Dorsal view of the skeleton of two oral angles and the first three vertebrae, x 12. Fig. 2. OpJdurochœta mixta. Dorsal view of the skeleton of two oral angles, x 10. Fig. 3. Ophiuroconis miliaria. Dorsal view of the skeleton of two oral angles, x 10. Fig. 4. Ophiaraclina incrassata. Dorsal view of the skeleton of two oral angles and the first two vertebrae, x 2. Fig. 5. BatJiypectinura gotoi. Dorsal view of the skeleton of two oral angles and the first two vertebrae, x 2|-. Fig. 6. Pectinura anchista. Dorsal view of the skeleton of two oral angles and the first two vertebrae, x 10. Fig. 7. Ophiaraclmella gorgonia. Dorsal view of the skeleton of two oral angles and one arm base, with the radial shields turned up, i.e. in ventral view, x 7. Fig. 8. Ophiochiton fastigatus. Dorsal view of the skeleton of two oral angles and one arm base. x4. Fig. 9. OpMoplax lamellosa. Dorsal view of the [skeleton of two oral angles and one arm base, x 20. Fig. 10. Ophiodoris pericalles. Dorsal view of the skeleton of two oral angles and one arm base. X 10. Fig. 11. Ditto. Dorsal view of two vertebrae near disk, x 10. Jour. Sei. Coll Vol. XXXVIII., Art. 2, PI. VI. Matsumoto del. Matsumoto : Japanese Ophiuroidea. H. MATSUMOTO. MONOGRAPH OF JAPANESE OPHIUROIDEA. PLATE VIL PLATE Vn. Fig. 1. Ophionereis annulata. Dorsal view of the skeleton of two oral angles and one arm base, x 10. Fig. 2. Ophionereis reticulata. Dorsal view of the skeleton of two oral angles. X 8. Fig. 3. Opiiiocrasis marldanneri. Dorsal view of the skeleton of two oral angles and one arm base, x 15. Fig. 4. Ophiocoma scolopendrina. Dorsal view of the skeleton of two oral angles and the first three vertebrae of a young specimen, x 15. Fig. 5. Ditto. Dorsal view of the skeleton of two oral angles and one arm base. x4. Fig. 6. OpJdomastix annulosa. Dorsal view of the skeleton of two oral angles and the first three vertebrae of a young specimen, x 10. Fig. 7. Opldarthrum elcgans. Dorsal view of the skeleton of two oral angles and one arm base, with the radial shields turned up, i.e. in ventral view, x 7. Fig. 8. OpMopsila riisei. Dorsal view of the skeleton of two oral angles and one arm base, with the radial shields turned up, i.e. in ventral view. X 14. Jour. Sei. Coll. Vol. XXXV I J I., Art. 2, PI, VII. Matsumoto del. Matsumoto : Japanese Opliiuroidea. Jour. Sei. Coll., Vol. XXXVIII., Art. 3. Prontispiece Sugiiiuito yiliotn. View of Saknra-jima in eruption, taken from the quay of Kag-oshima on Jan. 12th, 10.40 a.m., 1914. Koto : The Eruption of Sakuia-jima. nUNI'Kl) ]!V 'linC 'lOKYO riUNTlNO CO., LT». JOURNAL OF THE COLLEGE OF SCIENCE, TOKYO IMPERIAL UNIVERSITY, JAPAN. VOL. XXXVIII., ARTICLE 3. The Great Eruption of Sakura-jima in 1914. By Bundjirô KOTÔ, Ph. I)., RigakuhakusM, Professor of Geology, Science College, Imperial University of Tokyo. With 24 plates and 49 text-figures. Introduction. The eruption of the insular volcano of Sakura-jima on January 12th, 1914, was not only unusual in its magnitude, but also in its bilateral eruptions accompanied with copious outpouring of lava-flows which plunged themselves into the surrounding seas. As such remarkable phenomena seldom occur twice in a life-time even in our volcanic Japan, the writer made two trips to the actual scene of activit}^ Both the University and the Earthquake Investigation Committee granted a generous subsidy for the purpose. The present paper is nothing more than a preliminary record of what he has seen in the field and observed in the laboratory, and does not pretend to be an exhaustive account of the recent sub- terranean convulsions at the world-known Sakura-jima. Special acknowledgments are due to Assistant Professor M. Kanai, of the Higher School of Agriculture and Forestry, and 2 AKT. 3. — B. KOTÙ : Professors M. Uyeda and K. Yamaguchi, of the Female Normal School of Kagoshima, for their valuable assistance in collecting specimens of lavas and éjecta, and in securing photographs of Sakura-jima before and after the recent eruption. The writer is also indebted to Professor E. Junker, of the First Higher School here, for valuable suggestions during the preparation of the manu- script. Thanks are also owing to many other friends and col- leagues, whose assistance by their publications and in supplying information is acknowledged in the text. Part I. General Outline of the Geologic Structure of Southwest Japan. § I, Geographic and Geologic Situation of Southwest Japan. Among the insular garlands of Eastern Asia, there is an outwardly bending curve of island chain (Fig. 1) that stretches over Q^ degrees of latitude, and hangs at the north end of Taiwan and the southern terminus of Kyûshû, the latter l^eing one of the three islands that constitute " Old Japan." This linking chain is the E y ii k y ù a r c, which the writer lias long considered as the geographic and geologic homologue^^ of the Lesser Antilles with the classic volcanic islands of Martinique and 1) C. Brown acquainted ns with tlie geologic homologixe of the i:)lateaus of Tibet and the Shan States of Yimnan, l)oth being elevated ancient ocean-floors now undergoing abrasion and reduction to peneplains. The outer edge of each is bounded by a scarp exixising a zone of Archean and Paleozoic rocks which built up the Himalayan chain in the former, and the meridional western Yunnan ridge in the latter. In each case the zone is bounded by a great fault, making escarjjment and at the same time forming the inner edge of the fore-dee}) (the Ganges and Irrawaddy) that separates the edge referred to from the foreland (the Decoan and Upper Burma). Lastly, in front of the zone there occur faulted Tertiary strata (the Sub- Himalayan zone and the Tertiary of U]i]ier Burma). THlî GREAT ERUPTION OF SAKURA-JIMA IN 1914. F"ig. 1. — (îeotectonic Slnp of S.W. Jilp.'^n. ^ "" ^ .r %"^ifi/ C^ I l \^^ —— -^ /i:s' \ \ »30' ^^^^ SamboßeiwcL-MikaLn. Series CSchists) \^2^^_\ Piî^-Tunrj-hai Thiptiwc-LiixC' l^y "1 RyùhyvL Volcanic Chain. V'^\ Voltuxnic CJtain of Hot-nh- c^^iAciitR^ The writer may add the third member to the homologue, and that is the region iinder discussion. Indeed, the shallow Tung-hai is a submarine j)lateaii bounded on the southeast by a scarp — the cordillera of Eyûkyû, "which is also constructed of Paleozoic and green schists. This elevated zone makes the inner edge of the Ryùkyû fore-deep (700 m.) which separates the fore- land of the South Japan Sea, sometimes called Philippine Bay or ]\Iariana Sea — the siibmarine l^lateau of 400 m. enclosed by a large sweeping volcanic chain of the Bonin, Mariana and Palan islands. Earthquakes are frequent in the fore-deeps of the Himalaya, the Yixnnan-Burma frontier and Eyûkyû. The volcanic chain of Byûkyù is represented in Ujiper Burma by sporadic occurrences of volcanoes in the Teng-yueh (|§ '^) area. We have hitherto failed to find recent volcanoes at the foot of the Himalayas, although Hedin records their occurrence in the inner Tibetan plateau. .SVe C. Bro^^^l, Benord Geol. Sur v. India, Vol. XMII. Part 3, 1913. 4 ART. 3.^B. KOTÔ : St. Vincent. Like the corresponding Atlantic chain, the Ryûkyû arc is built up of three concentric zones,^ vi^., the outer Tertiary and post-Tertiary, the middle Paleozoic, and the inner volcanic. This ' volcanic chain of liyukyii ' begins witli the Togara chain (the Kawabé Seven islands, Linschoten or Cecilia Is.) and its southern extension (Fig. 1, A.), and proceeds northwards toward Kyushu through a number of active and dormant volcanoes to Kaimon (Fig. 1, Jfi ; Fig. 2, /ij, the last being the landmark and outpost of the inland volcanoes, located at the entrance of Kagoshima Bay. Sakura-jima, now under discussion, and also Kirishima should be coordinated with the already mentioned on the northerly prolongation of the same tectonic line, and likewise the central cones of the volcano of Aso in central Kyûshû. To go more into detail, the volcanic chain starts from (1) Aguni (Fig. 1, A-^, off the larger island of Okinawa. Then comes (2) the active Tori-shima T, which exploded in 1796, 1868, and April 11, 1903, on its northern prolongation. Among the Togara chain, which follows with an interruption of 130 km., on its northward extension dotted with twelve islands, we may only mention here (3) the konide of Akuseki A, , and (4) the burning Suwanosé S^. Of the latter we have the record of eruptions in 813, 1811, 1827, 1887, 1912, and lastly, on March 21, 1914, when it was spasmodically active every 20-30 minutes, emitting dense clouds, and the quaking of the ground was felt at Oshima at a distance of 150 km. The Meteorological Observatory of Kagoshima reported that from June 29, 1915, peculiar shocks were instrumentally recorded, especially on July 1st and 2nd. On the 6th, a Ryûkyû liner, the ' Okinawa Maru,' while coasting Suwanosé, noticed roaring and THE GREAT ERUPTION OF SARURA-JIMA IN 3914. 5 glare from a point at seven-tenths of the altitude on the eastern slope of the highest peak, which emitted strong smoke or ash- clouds to a height of 3-4,000 feet. Suwanosé was again strongly- shaken, producing fissures in the ground on September 3rd, 8 p. m., 1915, on account of which almost the whole population (250 souls) fled terror-stricken to Oshima. The volcanic earthquakes continued then for two weeks with waning energy. The active centre lies within the crater, 800 m. high, in the middle of the island. It remained tran- quil till the late strong earthquake, whence it slightly revived per- iodically ejecting ash-clouds. Suwanosé in sea and Kirishima on land {Sy and K in Fig. 1) are the two loci in the Ryùkyù volcanic chain which are at present constantly disturbed by volcanic paroxysms. Then come in the series (5) the imposing solfataric Nakano-shima N, and (6) the smoking Kuchino-shima K-^ in Fig. 1. Farther north there lies (7) the solfataric Kuchino-Erabu or ' Front Erabu ' K, in Figs. 1 and 2. Geologically speaking, the last is already situated within the shattered region {sec footnote p. 21 and p. 22) of southern Kyiishii. Still farther nortli we find the geologically connate equatorial scries of three islands, of which (8) Iwô-jima (7 in Figs. 1 and 2) or ' Sulphur Island ' — the Lipari island of South Japan — is active; it exploded on March 31, 1914. Iwô-jima was colonized seven hundred years ago. In annals of old families some records of volcanic activity may be found, if search is made {see p. 21). Beyond the Strait of Osumi or Van Diemen Strait on the north, we come to Kaimon and Sakura-jima as far as Aso at an average distance of 45 km. from one another. See Fig. 2, K^, S, A. A) The Peri-Tunghai Tectonic Line and the Genesis of the RyùkyCi Arc— As it has been already stated that Sakura-jima is b AET. d. — B. KOTO : one of the members of the ' Eyûkyû volcanic chain,' it will be well here to give a general outline on the geological significance and the genesis of the Eyûkyû arc. There is a tectonic line of first magnitude in Eastern Asia which, starting from the east coast of Korea, apparently terminates at the south end of Taiwan, thus enclosing the entire region of the shallow Tunghai or Eastern China Sea (Fig. 1). The writer will hereafter call it the Peri-lhcughai tectonic line. This rupture line assumes various phases on its long sweep. n) On the east coast of Korea it appears as step- East Ooast or KoEEA faults dropping eastward under the deep bottom of the Japan Sea.^^ Faulting OP South h) Ou its soutliward course it cuts transversely Japan Axis right through the solid axis of the mountain fold'^ which constitutes southern Japan, causing the west wing by this time to sink down to the continental shelf of the Tunghai. The shattered coast of western Kyûshû, the ' ^gean Sea of Japan,' owes its formation to this cause. The large crooked indentation of Ariaké, lying to the east of Nagasaki, is, geologically speaking, a part of the easterly lying Inland Sea or Seto-iichi that lies between the contracting end of Honshu and the island of Shikoku (Fig. 1). In western Kyûshû we see a peculiar but highly important crustal movement of the Peri-Tunghai rupture-line, which can be adduced from the orientation of the Smnbagawa-Mikahn Series'^^ 1) B. Koto, 'Journey tlirough Southern Korea.' This Journal, ^'ol. XXVI. Art. 2, 1909, Tokyo, p. 3. The faults produced are of the nature of what the liussian geologists call the disjunctive, and of the late von Kiehthofen's Zen-ungsver-werfung. 2) It is what geotectonists call the Blatt, or heave-fault. 3) The series is an old prasiuite complex of chlorite and graphite schists and overlying limestone, amphibole-schists, and clasto-pyroxenite of gabbroic derivatives, associated with glau- cophane and piedmontite schists. In the writer's opinion, the series rexire.sents a metamorphic THE GREAT ERUPTION OF SAKURA-.TIMA IN 1914. 7 (Si', ill Fig. 2). The series, shaded black in Fig. 1, conforms to the curvature of South Japan in detached bands. In northern Kyiishû it widens and suddenly deflects southward near Nagasaki, continuing farther southward along the west coast of the islands of Amakusa. After an interruption of four degrees it reappears unexpectedly in Oshima, whence it follows the curviforni structure of the Eyûkyû arc {><('e Fig. 1). cMN^ ^) ^^^^^ Ryûkyû arc is monoclinar^ in structure, ^Ïrc^ dipping westward, and facing its tilted edge toward the fore-deep of the Eyûkyû Graben (7,000 m.) on the open Pacific (Fig. 1). The overthrusting movement of crust-wave upon the fore-deep, accompanied perhaps contrariwise with underthrust, seems to the writer to be the cause of the curvilinear alignment of the sub- marine Cordillera, whose crests appear as the serial islands of Eyûkyû. A reactionary regressive movement on the continental side caused depression (500-1,000 />?.) and loosened the crust within the continental shelf — the Asiatic bank, giving occasion later to the outburst of magma. It is the inner zone and at the same time the Volcanic Chain of Eyûkyû. (Fig. 1) The Peri-Tunghai tectonic movement happened between the Cretaceous and pre-Lepidocyclina Tertiary times.'^ From the main distribution of typical marine Tertiary beds which fringe the outer side of the cordillera and also Taiwan, it seems probable that the faciès of the whole Lower Paleozoiciim, especially of the Devonian. As it is the oldest complex in our islands, excepting the basement gneiss which is of limited occurrence, the writer has chosen that as the exponent of the stratified bed. 1) It is usually spoken of as a portion of folded mountains which constitute the axis of the southern wing of the main Ja^janese islands. So far as the writer knows at present, there are no folded mountains in existence, biit that of isoclinal flexure-xsmQes which simulate the folded structure. 2) Prof, l'abé informs me on a siiecimen which I sent him that a Niimmulite, Pellastispira, occurs in the limestone from Motobu in Okinawa Island. He assigned the age of the rock to be either the Ohgocene or the Eocene. s ART. 3.— B. KOTO : principal form-casting tectonic disturbance occurred before tho deposition of the Tertiary complex in question. The faulting and other vertical movements, however, took place in the later Tertiary, or early Quaternary period, which movements still continue down to the present, as is evinced by frequenter earthquakes. This geodynamic history is in conformation with that of Central Asia and peripheral China.'^ The volcanic activity of the Kyûkyû chain began concomitantly with the above-mentioned later move- ments after the deposition of the thick lapilli bed, and the for- mation of the trench bay of Kagoshima. It will be referred to again in the sequel (p. 17). The volcanic chain, as is frequently the case, traverses indifferently and in the present case obliquely, the ground structure of the Kyûkyû arc, and enters right through the middle of southern Kyûshû, tSakura-jima being one of the vents upon the chain. An inspection of the bathymétrie conditions in Sketch-map Fig. 1, can tell the facts better than the writer. Located between the two parallel deeps — the Ryûkyû fore -deep (Graben) and the inner rift- valley (Fig. 1), the Ryûkyû cordillera might have suffered ' framed folding ' (Suess's Rahmen faltung). It is an open question whether the present volcanic chain stands on the margin of ' relaxed ' horst, here meant by the Ryûkyû cordillera, as in Sumatra,'^ or not ; but at the distal end of the chain in southern Kyûshû several well-known volcanoes lie within a negative land-form — of which the trench depression, Kagoshima Bay, gives us unequivocal evidence (Fig. 2). We have to speak about it later on. 1) Maclaatschek, 'Neuere Arbeiten zur Morphologie von Zentral- Asien.' Geogr. Zeltschr., 1914. 2) Volz, 'Die geomorphologische SteUimg Sumatras.' Geogr. Zeitschr., 15. Jahrg., 1909. S. 11. See also ' Die Gajoländer,' by the same author, Berliïi, 1912. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 9 It is contrary to modern views that here the volcano of Sakura-jima and many others sit upon the compressed, wedged-in bottom of a rift-valley instead of the loosened high edge of the same. Such examples are, however, not wanting in fissure-eruptions and effusions in Iceland. PBoI^Nor ^^) ^^^ passing, the writer may mention for the sake ™îûKY?*^ of completeness the southern extension of the outcurve of the Eyûkyû arc. Adhering still to the same principle, we trace the trend of the ancient, green schistose Sambagawa- Mikabu beds through a detached group of the Saki-jima Group toward the stupendous diorite cliff of the Dome Cape of Taiwan (Fig. 1). Just as ill the Kyùkyû arc, the fundamental structure of Taiwan remains the same, being homoclinal and also dipping westward. The same series in Taiwan, however, pursues a southward course, curving inward and westward ; consequently, the east side of the island corresponds to the inner zone, and indeed a volcanic chain runs along the inner, east side. See Fig. 1. In short, the guide line of the southwest islands of the Japanese territory describes an unbalanced sigmoidal course, the Byiikyii curve faces its convex, and the Taiwan curve^' its concave side to- 1) The geologic structure of Taiwan is unique. The festoon islands of Kastern Asia, as in aU similar cases of cur\ilinear island chains elsewhere, face their convex side toward an open ocean, while Taiwan alone (see Fig. 1) behaves contrariwise. The late von Richthof en in his oft-cited Morphologische Studien aus Ostasien called our attention to this point, and said of this island that it is geologically neutral, Avhile leaving the question open on the cause which produced this peculiar geologic structure. The writer also believes that Taiwan is indeed a neutral region, and has been at a standstill, while the Phihijpine and .Taiwanese islands disproportionately buhfed out toward the western arm of the Pacific, which is a gigantic depression on the ancient continental border of Eastern Asia, and which the writer habitually ciiUs the ' South Japan Sea.' As a result of a conference with ^Messrs. Yamané and Noda, of the Geological Survey, who recently made journeys through South China, the writer is impressed with the idea that the inclined tableland of Fo-lden, together ^^"ith the adjacent portions of Kwang-tung and Che-Mang, conforms to the incurve of Taiwan, -with tilted edge this time toward the interior of China. This is the all important tectonic line in southeastern China, which has hitherto escaped the observation of geologists. 10 ART. 3. — B. KOTO : ward the open Pacific. Through the sinuous point or middle wing (Fig. 1) between tlie positive and negative curves, which is located at the northeast point of Taiwan, a volcanic line passes from the inner side of the Eyiikyû outcurve to that of the Taiwan incurve. Essential ^a i^ sliouM, howcvor, bo bomo lu miud that the Chaeactee- ^ ^"^RïûIïT^ volcanic chain of cantalitic hornblende-andesite of Tai- VOLCANIC r,, . . j_' i? 1 1 Chain wau rcappcars, alter an mterruption oi seven degrees, at Unzen and Tara- daké (?7and T in Fig. 2), both near Nagasaki, and then proceeds through the heart of Kyûsliû at Kibo-zan B, Kujii K^, Tsurumi and Yufu-daké Y, and Futago-yama F, and finally passes over to the contracted end of Honshu along the shore of the Japan Sea (Fig. 1). TJic Ryuhfii volcanic chain, cliaractertzed hy augitc- andésite or auganitc, on flic other hand, only starts froin Agnni {A. in Fig. 1) off Okinawa, and, as is frequently mf^ntioned, runs through tJie Togara group to soutliern Kyushu. We have, therefore, to distinguish two volcanic chains quite different in the nature of magma, viz., the cantalitic hornblende- andesite and the femic augite- andésite. To the latter group be- longs the Kyûkyû volcanic chain ; and Sakura-jima, which is the subject of the present paper, is one of the members of this chain. B) The Principal Formations and Main Geologic Structure of Southwest Japan.— From a glance at the map of the southern, portions (Figs. 1 and 2) of both Kyùshû and Shikoku — the two main islands of " Old Japan," one can readily understand the fundamental geologic structure. The principal geologic formations appear in parallel zones, which cross them from the northeast to the southwest, and the two islands are the //o/'sf-blocks in the geological sense of the term, with the rias coast on the eastern and western shores. Tin-: GItEAT F-lUTT OS OF SAKUlîA-JIMA IX 1914. 11 o) The iiortluriiiiiost '/.vno is bnilt u|) of the oldest foi'iUHtioiis of Gneiss {Ar hi Fig-. 2), variously intruded and assimilated by granitio magma. The region is for a greater part occupied by the equatorial Inland Sea or Seto-uchi, and the K^aishû portion of it is hoav hiddtai mider the superimposed ^"olcauic masses.^-' h) The next in the order of age Hes to the south of the first zone in a narrow contracting band ; and in northern Kyîisliû it reappears on the north side of the first zone. It is tlu; green schistose Samba- gawa-Mikabu complex (Sc in Fig. 2) of earher Paleozoic age already' refeiTed to.-^ e) Small patches of Cretaceous bed Cr occupy a narrow space between a and b zones on the east ; Ijut on the A\(^st coast of Kyûshù it covers the submerged Jans: ça —SnijTu^ YoJcti^S area of the Gneiss zone in the Amakusa islands at the south of Nagasaki. d) Next comes on the south side a broad band of Paleozoic graywacke formation P<(, devoid of fossils. é) The southernmost zone is the sandy slate formation M of probably Mesozoic age. It is this formation which covers the three provinces of Fig. 2. — Geologic sketch-map of Southern Kyûshù scale , 1 : 2,000,000 Volcanots : /I'l-Kaimon, S-Salaira-jima, ^T-Kirishima, ^-Aso, Jifi-Kujû-san, F-Yufu and Tsurumi, F-Futago, ^-Kibô- zan near Kumamoto, L'-Unzen, T-Tara-daké. Formations : (r-Granite, Ar-Archean, Sc-Sambagawa-ililcabu Schists, Pa-Paleozoic, Cr-Cretaceoiis, il-lMesozoic, T- Tertiary, La-Platcau Formation. 1) These surficial volcanic masses are left out in Fig. 2. 2) See page 6, footnote (3). 12 ART. 'S. B. KOTU : HjTÎga, Ösuini, aud Satsuma, comprising au area of about oue-tliird of Avliole Kyûsliû (39,659 sq, 7cm.). Poor seams of authracite are iuter- calated iu this complex. Both the northerly lying Paleozoicum Pa aud the southerly lying Mesozoicum M in Fig. 2 — the two zones d and e — dip together toward the nortJaoest, and on this account their geological relations are not at all clear to the writer. Topogi'aphically speaking, the boundary of both zones is the liighest elevation (1,820 m.) and water-shed, traversing obliquely the whole breadth of Kyushu. Granitic masses G occur ^vithiu both terranes, one at the southernmost end of Kyushu, the other at Yaku-shima G. f) Tlie Tertiary bed occupies only a small patch on tlie east coast; but it builds up the whole island of Tanega-shima (jT in Fig. 2). It is characterized by the presence of OpercuUna, just like the petroliferous Tertiary of central Taiwan. In short, the three principal formations a {Ar), d (Pa) and e (31), counting from the north and in the order of age, traverse obliquely the geologic Jiorst of the Kyûshû block (Fig. 2). g) Having given an epitome of the principal formations and main geologic structm'e, the writer has still to mention a surficial bed on the Mesozoic (M) and Tertiary (T) terranes, which bed has a great bearing in deciphering the mode of formation of Kagoshima Bay. The land sm-rounding the bay is a plateau of lapUli and ash, which attains an altitude of 230 m., and which bounds the shore in sharp perpendicular clifis, as if the bay were a sunken gigantic caldera, for which indeed it was taken by the late F. v. Richthofen. It is the Plateau Formation (La). The bed occupies a greater or lesser portion ot the provinces of Satsuma, Osumi aud Hyûga, comprising an area of 1,G50 sq. km. From the terrene rise the Mesozoic slate formation M and granitic masses G in Fig. 2, and it is underlaid aud in part over- laid by volcanic masses, one of the latter being the well-known volcano of Kirishima. THE GREAT EEUrTION OF SAKUEA-JIMA IN 1914. 13 § II. On the Formation of the Trench Bay of Kagoshima. A) General Geology of the Environs of Kagoshima Bay.— As to the outstanding question about the foundation, on which the volcano Sakura-jima stands, it is first of all necessary to know the geology of the border land of the bay. As has just been stated, the extensive (1,650 sq. km.) and thick (100 ??2.), ash-grey and loose bed of lapilli and ash is a characteristic formation of the present area, forming a high flat of more than 230 m. Tliis formation is underlaid by a complex of sandy slate of problematic Mesozoic age, and the Tertiary Operculma' and shell-bearing sand- stones with coaly shale. Both are intruded at several places by the ash-stone which plays a great rôle in the geology of the region. Ash-stone, a) The jish-stoue^^ or hcd-ishi — îi hypersthene-tracliy- Laya andésite (PL XYI. Fig. 1) is an ash-gray, friable and porons rock, so named on account of its having the appearance of hardened volcanic ash. It is extensively quarried for building materials, due chiefly to its being easily worked rather than to its intrinsic value. The a variety is dopatic, with phenocrysts of idiomorphic oligoclase and corroded sanidine. Femic minerals, such as hypei*stliene, augite, hornblende and biotite, are found only sparingly and often not discovered in a shde, the first component being, however, comparatively abundant. The main bulk of the rock is devitrified glass of various degrees of alteration, often spherulitic (PL XVI. Fig. 2) and sometimes silicified. The ß variety is a black porphyiitic obsidian, and sempatic with macrophenocrysts of feldspars, the black base being colorless glass swarming with streams of feldspar-microlites. It seems to correspond in a strict sense to the so-caUed Aso lava. (PL XVI. Fig. 3) The light-grey y (PL XVI. Fig. 2) modification makes shreds in the 1) It is a pipemo-like eutaxitic rock, striped with black glass in j)umiceous white mass. The stripes may be either schlieren-like patches detached by flows, or doughs of lava drojDped in white ash mass. It is habitually called the ' Aso lava ' by our geologists. 14 ART. 3. 13. KOTO : preceding, from wliicli this one differs only in tlie groimdmass, being mainly splierulitic, and gradually passing to splierulitefels. It frequently makes an independent mass. It is this ash-stone which seems to the writer to be the same effusive, named plagioliparite or rli^'odacite, found in ])atches all over Japan, although here quartz is not typically OKiGiNor ,-ie^-eiopeci. It is this rock which gave birth to the oft- mentioned sulmiarine bed of lapilli that is usually misinterpreted as ] laving been ejected from either Ivirishima or Sakura-jima. In order to make a close study of the geologic data given above, as well as to illustrate the mode of ^ Cap-sheet cf Lava fhty ASH-STONK AND Lapilli Bed in the Enttrons OF THF OtTY OF Kago- formation of the bay itself, the writer will select tlio SHIMA environs of the city of Kagosliima for this scope. The city lies opposite and west of the volcanic island of Sakura-jima, and is separated from it by a channel 3.7 h.m. wide and 25 fathoms deep. The old castle- hill of the Feudal Lord Shimazu stands in the city on the plateau- edge, 100 m. high. It is the fjipe of the lapilli plateau of southern Kyùshù. A. — The upper massive, 80- meter bed is built up of an incoherent ash-grey mass of pumice and lapilh', loosely cemented with ash, under- laid (at the entrance of the railway tunnel) by B. — A liglit-l)rownish clay bed (Fig. 3). It consists mainly of ^. ^ . ^ comminuted viumice and a few clay- Fig. 3.— Section of L.-xpilli rktean near Kagosliima. <^y particles with some In'persthene THE GIŒAT EKUPTION OF SAKUEA-JIMA IX 1914. 15 inul liornbleude ciystals. TJie fossils found by the ^vriter ;ire Dcntalinm sp. Hud il nin\alv(\ This oürrespouds to the fossil hori/on at Kaikata iininediately to be reforred to. The two beds A and B lie conformably with <>TndnMl transitions. C. — The bottom exposed here is of a gre}', soft and soaking mass of liai-isld or hypersthene-trachj'andesite, already briefly outlined (p. 13). This is a sul)mariue lava-flow wliicli gave birth to the lapilli. PI. XYI. Fig. 1. Following the railway from Kagoshima northward along the shore, the effiisive C becomes sphenüitic with pjT-'oxenes at Jigami Station, i.e., the r tyi)e (p. 13). PI. XYI. Fig. 2. At Shigetomi the lapilli plateau, 500 m. high, is capped with trachybasalt A^. PL XVI. Fig. 4. Farther northward, at the foot of Kirishima, the cap-rock is ji hornblende-bearing pyi'oxene-andesite A-^, which apparently attains a thickness of more than 200 III. {see Fig. 3). The extensive lied of l)lock-mud capped with liornblende-p}TOxeue- andesite in northern Kj^ûshû, especially in the Yabakei region in Prov. Bungo, seems to the writer to be the equivalent of (jur lapilli bed ^i and the cap-rock ^L. The Yalmkei region is well known for its canyon scener3^ The lapilli bed or its equivalent is, therefore, regional, and by no means confined to a corner of southern Kyfishû. The stretch, 100 to 500 ni. high, between Kagoshima and Shigetomi just referred to, and all the pla.teau-lands (400 m.) on the southern foot of Kirishima di^scend precipitoush', a/., :d Kokubu Statiou,^^ to the shore-flat, or directly to tlie water's (;dge (PL lY. Fig. 1) — a special 1) After having finished the manuscript, the writer came across the passages in -which the lite Y. Pdchthofen has not only given an exact written profile of hornblende-bearing l^umiceous complex capped with trachytie rock, accompanied with pétrographie details at the north of Kokubu, but also the canyon-like erosion valleys in the jilateau land, which characterize the scenery of this region. (' Geomorphologische Studien', III., S. 16.) The writer was greatly surprised at the author's keen observation many decades ago, \\hen that part of Japan was still entirely unknown ; but his comparison of the Bay of Kagoshima and the Sakura-jima volcano with the Laguna de Taal in Luzon seems, however, not a fortunate one. The writer considers the bay as a rift valley and not a caldron depression ; but it nmst be remembered that a partial nature of Icottle depres.sion is not yet fundamentally confuted. IG AKT. 3.— B. KOTÔ : cleau-cut topography, ^vliicli eloquently speaks for tlie depression of the whole complex to the bottom of 100 fathoms of Kagoshima Bay. Eegion°on ^^^*^ same geologic feature can be seen on the posiTE Coast Opposite coast near Seto Strait, noAy entirely buried and land-locked by laya-floAVS of the last eruption. The table-mountain of Sakkabira (Ho-zaki) (325 on., PI. IX. Figs. 2 and 3 ; PL X. Figs. 1 and 2 ; also Geologic Map), facing the strait, is again the ash-stone or piperno C (Fig. 3) of rather compact texture and light-gray tint, capped with a dark-gi*ay trachybasalt Ai in Fig. 3. In the former, i.e., the ash-stone or liypersthene-trachyaudesite, hornblende or biotite are occasionally found as accessories. The ß and f varieties (p. 13) are exposed on the shore near Obama, 1 hn. south of the strait, o in PI. XII. Fig. 1. Land- ward we find the Mesozoic slate, striking N.E. to S.W. with northwesterly dips, in contact with granite batholith. See Geologic Map. Both are intruded b}' the ash-stone C, which was eroded later together with the slate and granite mass, resulting in the formation of flat- topped mountains, 300-700 m. high. They are flanked on the western shore by the 200-meter terrace of lapilli bed. (PL XII. Fig. 1) At the fishery village Kaikata, south of the above-mentioned Obama, the ash-stone is overlain discordantly on the eroded edge by a gi'avel bed of slate, ash-stone and ash, which corresponds to the horizon B of Ivagoshima (Fig. 3, p. 14). The writer made here a collection of a number of fossils which were sent to Prof. H. YabÉ for examination. The fossils determined are Gardium iimticwii E,Ave.,. C refusion L., Dodnia troscheli Lischke, Tapes euglypbis Phil., Bulla vernicosa var. ovula ?, Cassis sp. indet. Yabé remarked that all are living species. No. 2 however, being rarely found as a fossil, and that the age of the bed B in question may be of either the late Tertiary or the Diluvial. Another locality, made known by M. Uyéda, is 0.2-0.3 hni. inland from the shore t)f Tarumi, Avliich lies to the south of the aforementioned Kaikata. The sandy fossil bed underlies the siu'face k)am deposit, 1-1^ 711. thick, and is probably of the same age as the Kaikata bed. The THE GREAT ERUPTION OF SAKURA-JBIA IN 1914. 17 l)ocI is full (»f Gardiuni (Liiunlicai'diuni) rctusrim L. aud PJtos senticosus L., wliicli are Ixnug used as raw materials for limekilns. From the attitude of this bed toward the Tertiary Operculina sand- stone in another region (near Miyazaki', the writer is disposed to consider with Yabé the fossiliferous gravel bed to be of the later Tertiary or the eay'hj Diluvial age. The fossihferous layer is seen at the above- mentioned Kaikata, often intercalating with the lapihi bed, 100 m. thick, wliich contains blocks of ash-stone. The lapilli and fossiliferous horizons {A and B in Fig. 3) are not separable and both are submarine deposits. EAi*i^">f We are, therefore, forced to the conclusion that the Age or entire Plateau Formation, now elevated 200 m., must be Lapilli Plateau of later Tertiary, or more probably early Diluvial ageP As the plateau edge, 100 ??2. high, forms cliffs all along the shore as on the Kagoshima side, we have again here the proof ^nn a^^5ay" ^^^^^ KafjosMma Bay is a Jiollow in the lapilli formation, FoKMH,D depressed during the Diluvial period. Considering the average elevation of the Plateau-land to be 230 m. and the bottom of the bay 200 ?j^l or 100 fathoms, the vertical relative displacement of the crust will therefore amount to 430 m.— di, considerable crustal displacement which occurred in a comparatively recent geological period in southern Kyûshû. Having dwelt upon the region of Kagoshima City, and of the defunct Seto Strait, the writer will now shortly touch on the third region of the lapilli land, where the crucial test can be advanced to the actual dislocation by wliich Kagoshima Bay was created. Geology At the cntrancc of the Bay of Kagoshima stands AND DlS- IX5CATIONOF tho promineut kouidc (see Fig-. 2) of Kaimon, from which theKaimon ^ \ o / ^^^ the view (PI. IV. Fig. 2) was taken northward to show 1) The Plateau Formation may perhaps represent both the late Tertiary and the early Dilmiuin without unconformity between them, on the supposition that the physiographic environment remained the same throughout the whole range of time. 18 ART. 3. — B. KOÏÔ : the escarpment of dislocation on the left side. The region is likewise built up of ash- stone C, or hypersthene-trachyandesite apparently with some quartz, overlaid by block-mud B, and then lapilli deposit A {see Fig. 3). The rolhng surface, 200 m, or more high, built up of these rocks, descends abruptly eastward, as in the figure, by faulting toward the lowland of Kaimon, on which at later times the konide of the same name and the gigantic crater-lake of Ikéda (PL IV. Fig, 2) with a diameter of 4 hn. made their appearance, the one being the positive and the other the negative form of nearly equal size and volume.^^ The re- markably regular escarpment of over 100 ??î., so produced, runs with a trend to N.N.E. as in the figure. It is a part of the cliffs that bound the western shore of Kagoshima Bay. coxcLtTDiNo What has alreadv been stated will be sufficient xlEMAEKS ON '' M^noN^oF ^'^^ the present purpose in regard to the geology, and Bay the way which led directly to the formation of Kago- shima Bay. The writer will now summarize the facts mentioned and the suggestions made before. The bay is a slightly- bent trough at the south end of, geo- logically speaking, the oblique horst of Kyûshù, and at its northern blind recess rises the volcano island of Sakura-jima. The bay is in average 20 km. wide, and 60 km. in its axial direction as the prolongation of tlie volcano chain of Eyûkû. That it is a topographic hollow was long since stated by v. Eichthofen and SuESs, from tlie morphological aspect which it presents. Conscious of this or not, our geologists hold the same view ; and especially since the recent paroxysmal convulsion of Sakura-jima, the super- 1) The diameters of the Lake Ikeda and the base of Kaimon are about 3.5 km., and the volume of the lake also apiiroximately coincides with that of the volcano, being about 4 cub. Inn. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 19 abundant literature on the eruption by specialists as well as by laymen is diffused with phrases of the same tone. But so far as the writer's knowledge goes, no one has ever studied seriously the geology, the mode of formation and the age of tectonic move- ment in the bay. For this reason, the writer has indulged some- what in detail on the above subjects. Epitome rpj-^^ geology of the rogiou, briefly told, is this : OF TSEj Geology rj,]^^ regiou arouud Kagoshiuia Bay is built up of Mesozoic slate and batholithic granite. On the eroded surface of submerged land deposited locally the Tertiary formation, whicli was later partially upheaved and eroded. Then \Tilcanism was called into play in the shallow sea, when the ash-stone C (Fig. 3) poured out, catching fragments of slate, which solidified under water as soaking stone and spherulitefels. Then sands and gravel B, products of erosion, deposited with marine shells. Vulcanism was still in full force, ejecting submarine pumice and lapilli A, which were deposited to a great tliickness of 100 m. or more. Meanwhile the sea was sinking while deposition was going on. Finally, the land was upheaved to 230 m. forming a monotonous lapilli plateau which characterizes the region (PL IV. Fig. 1). The long interval of vulcanism corresponds to the later Tertiary or early DUuülal tlms, which we know by the evidence of fossils. Cotemporaneous with that upheaval, vertical dis- placement of 430 m. took place, thereby creating the trough-bay of Kagoshima. The whole coastline of Kyiishii was remodeled during this period. It is contrary to an assertion often made recently that here volcanoes find their vents in the sunken bottom, as Sakura-jima and Kaimon, instead of the highland margin of the tectonic graben. The writer is inclined to think that volcanic vents find 20 ART. 3. — E. KOTü : their orifices only on the edge of a rift-valley, ichen iJie depressed and compressed crust makes uudertlirust against the stand'mg block, which in consequence rises and expands and also hursts open, therehij giving occasion io the escape of magma. In the present area, the undisturbed plateau simply sank down clear-cut in trench form, the bottom of which w^as, how- ever, multifariously shattered into polygonal blocks, which appear above the water of the bay as the islets of Oko-shima, Hakama- goshi, Ko-jima, Kanzé, and Karasu-jima, the last being buried under the recent lava-flows. Simple fragmentation is directly opposed to the compacting of the wedged-in crust. This may afford some opportunity to the upward stoping and the eruption of active magma. Volcanoes make their ai^pearance in a diffuse manner in such a pattern of great magnitude. In another aspect it seems to accord well with Reck' s second case^^ of the occurrence of eruptive vents, that is to say, within a trench valley the out- bursts follow the fracture-hue which bounds the table-horsts. In short, the plateau of Satsuina and Osuinl is a table-bloch, while the table-fracture tcithin the block creaied Kagoshima Bay, and the fragmentation of dropped block gave occasion for the escape of magma to the surface. B) The Outlying Ösumi Group-)— Tlie extensive plateau-land of lapilli is not confined to the above area. It occurs not only in detached patches in central- western Kyùshû, but also on the southern extension of Osumi and Satsuma in the form of inland groups, where it occupies a total area more than that already mentioned. The writer adduced the facts of this unexpected dis- 1) Wolff, 'Der Vuliiiuismus,' I. S. 111. 2) See ante, p. 5. THE GREAT ERUi'TIOX OF SAKURA-JDIA IN 1914. 21 coveiy from the keen stratigraphie observations made by I. Friedlaender, although he seems to be unconscious of the real extension of the lapilli plateau land. Through the effort of the German naturalist,^^ some light is thrown on this vulcanologically little known group of Iwô-jima and Kuchino-Erabu. He speaks of Iwo-jdia (Fig. 1, /; Fig. 2) as a high flat of 100-200 m., being built up of vitroandesite overlaid by tufaceous 1) ' Ueber einige japanisclie VuUv.ine.' Mitleil. Deutsch. Ges. f. iSfahir- u. Völkerkunde Ostaslens. Bd. Xn. Teil 2, Tokyo, 1910. For the sake of comparison and the cognate relation which they bear with Halmra-jima, a few notes will be appended chiefly from the work cited above, on the islands now under con- sideration (see Fig. 2). a) The active Iwo-jima (' Sulphur Island ') lies 18 miles from Cape Kaimon, and has a circumference of 12 km., the western half is a high flat, 200 m. in height, being built up of vitrodacite, tuff-sandstone and pumice, capped with lava-sheet. The eastern half is occupied by an active cone, 770 jn. high, which rests upon the pumice terrace. The cone carries a top- crater having a diameter of 7-800 m. and a depth of 50. Fumaroles and solfataras are hissin» from the bottom. The island is the well-known producer of sul^jhur and supjjorts 900 in- habitants of whom one-third find their employment in working the sulj^hur deposits. We loiow of no historic record of any eruption. As it was colonized seven hundi-ed years a^o, old families must keep some records of activity, which may be found if search is made. At the beginning of April, 1914, it showed an unusual mood of vulcanism in throwing up smoke in sympathy with Sakura-jima (See p. 5). h) The smaller Takeno-shima lies to the east of it. It is a flat island bounded on all sides by cliffs. The geological structure is exactly the same as the foundation of the preceding. Friedlaender imagines the island to be a part of a ruined cone whose center lies in the bottom of the southern coast. To the writer it seems more plausible to take the island for a patch of shattered lapilli land. We entirely miss i^etrographic details. Younger cones are not found on the island. o) The westerly-lying 1vubo-shim.i is a 'black' wooded conical island of the size of the preceding. It is said (' Geographic Scraps of Satsuma, Osumi, and Hyûga ') that to reach the steep top (250 m.) is by no means easy, and there are cataracts within the thick forest on the way up to the crater which seems to be of recent origin. We know nothing about its rocks and eruptions. d) Kuchixo-Ee.abu lies not far from Yaku (see p. 5). It is ill-defined, stretching VI km. from N.N.W. to S.S.E. According to Friedlaender, it is likewise an elevated flat (150 ju.) and cliff-bound, and has the same lithologie elements and structure as IwG-jima and others. It .•arries upon the foundation the overlapping cones (G65m.), the north cone being older and craterless, the southern one still preserving a fresh solfataric crater. At the middle stands the Latest-bom giant with an apical crater, 500 m. in diameter and 50 m. in depth, where strong fumaroles are hissing and depositing sulphur. A hot spring gushes at the northeast coast. A violent eruption is said to have occurred in 1839, when many people were buried alive. At the end of [March, 1914, it Avas rumored that it had thrown up thick s)noke, which was observed from the easterly-lying Yakn-shima, in .sympathy with Sakura-jima. 22 ABT. 3. — B. KOTü : sandstone and a strong bed of puiniee, capped with a sheet of lava-flow. The eastern sister island, Takeno-shima, has tlie same topography (100 m.) and the same lithologie elements. KuEO-SHiMA, lying 17 miles west of the former, was seen by him, however, only from a distance, and we know nothing about its rocks and eruptions. Probably it is of the same type as the two preceding ones. Kuchixo-Eeabu (Fig. 2 ; Fig. 1 K.^ is also a clifF- bound flat island, 150 in. high, and geologically and topographi- caUy an exact copy of Iwô-jima, above referred to. These four islands were later crowned with recent volcanoes. CoxcLTJsioN From the similarity in the geology and topography of adjoining lands and the insular group in question, we can picture in our mind the region to have been once an extensive plateau - land, which was later shattered, contemporaneous with the creation of Kagoshima Bay, Oshima Strait, and the intervening seas between the shattered islands. If the writer is not mistaken in Jtis suggestion, the îcliole group of the ahore-mentioned islands, together with Yahu, Tankga-shima, Mage, and Kiichino-Erahu, must have once formed an integral j}art^^ of ancient southern Ki/ilshit. C) The Foundation of Sakura-jima. — The question as to the immediate foundation of any volcano cannot usually 1)0 easily ans- wered. In a case hke the present, wliich is that of an isolated insular 1) See the area clotted with islands in Sketch Map of southern Kyushu in Fig. 2, and also the region enclosed by the 500-meter isobathe ofE southern Kyûshû including K-i (Kuchino-Erabu Island) and I (lAvô-jima) in Fig. 1. The conclusion arrived at here is also confirmed from the Zoogeographie point. From the study of the Japanese termites, of which now twelve in number are known, I'rof. S. Watasé says that the two islands of Yaloi-shima and Tanéga-shima form the southwestern outpost of the Pala-arctic section of the Japanese Empire, while the island of (Jshima (Anami-Ûshima) marks off the extreme northeastern boundary of the Oriental Kegion. See Figs. 1 and 2. The southern border of the shattered region of southern Kyûshû, above referred to, is therefore an important faunastic divisional line of old geologic date from a termitologic point. See X. Holmgren, ' Die Termiten Japans.' Annotntiones Zoolorjica; Japonenses, Tôlcyû, ^'ol. VIII. Part I., 1912, footnote p. 10!). THE GEEAT ERUPTION OF SAKUEA-JIMA IN 1914. 23 volcano, wo have first of {ill to consider tlie mainland that lies close to it, and this the writer believes to have done in the preceding pages (pp. 13-20) in connection with the discussion on the geology and the origin of Kagoshima Bay in which Sakura-jima is situated. As already remarked, the bottom of the trench bay of Kago- shima was multifariously slmttered (p. 20), leaving some polygonal blocks as if they were wedged out from the rest of the members. The tectonic movement was therefore similar in kind to that, by which plateaus are usually produced. These uplifted blocks are the low flat islets^^ of Oko-jima (40 m.)p Hakama-goshi (72 y«.),''^ the low Kanzé, the lava-drowned Karasu-jima,^^ and Ko-jima,^^ the last- named being at tiie nortli end of the bay. The bottom is 100 fathoms deep, and is a part of the continental shelf. The geology of these islets is interesting in many ways ; firstly, they were never closely studied before, and secondly, they give some clues as to the nature of the foundation of S a k u r a - j i m a. Oko-shijia 1) Oki-kojima (spokcu Ohosliimci) lies on the south- west coast of Sakura-jima (Geologic Map ; PL III. Fig. 1), near Moyé-zaki (the ' burning cape ') — the lava-end of the eruption in 1475-147G. It is only 400 m. long by 200 m. broad. It rises from the bottom (10 fathoms deep, and its flat top (40 m.) is GO m. lower than that of the ancient citadel of Kagoshima. It is said''^ to be a part of the lava-flow of 147(), which was later severed fi-om the land by a depression of the intervening sea bottom ; but this story is entirely unfounded (p. 42). 1) The new isLands off the northeastern shore of Siilair.i-jimn, formed during und after the eniption of 1779, will be dealt with elsewhere {]}. d8), since these historic islands form a special group by themselves. •2) i^ ^h h !^e Geologic Map. 3) :ff fg or Shiro-yama (Mc lij) 4) .0? % 5) /J. ^ It lies outside the limit of the annexed Geologic IMap. 6) 'Compendium of Geography of Satsuma, Ôsumi and Hyûga,' Kagoshima. ^ P^ H üi ?1 24 ART. 3. — B. KOTJ This gunboat-shaped islet'' (Fig. 4) presents the characteristic feature of the plateau land of Kagoshima, or a marine butte, and it is really a fragmentary block of the latter. Geologically, it is an exact copy of the Kagoshima environs (p. 14), the lower two- thirds being built up of the ash-stone or biotite- bearing hypersthene- Fig. 4.— Oki-Kojima vie wed southwestward from Sakura-jima. O. — Oki-Kojima, cm. — Cape Sloye-zaki. trachyandesite"^ (PI. XVI. Fig. 1) C in Fig. 3, overlaid discordantly by a thin horizontal, water-bearing bed of yellow sandy tuffite which corresponds to the fossil horizon B, and lastly, the lapilli bed A which makes up the upper one -third of the vertical section. One can readily recognize these beds even from a distance through differential erosion (Fig. 4). The andésite shows flo wage-bandings which dip at high angles in different directions. The normal trachyandesite is grey and light ; the grey perlitic ß type with hornblende (PL XVI. Fig. 3), and the black porphyritic obsidian type y (PL XVI. Fig. 2) with pyroxenes (p. 13) also occur with pillow structure, indicating collectively that the ancient lavas are of submarine flows. At the north end of the islet a dark-gray 1) In 1864, it was fortified for defense against an English squadron Avhich made an unsuccessful attempt of bombarding the city of Kagoshima. 2) Some contain a large quantity of corroded quartz, and in this ease the rock may be appropriately called dacite or plagioliparite. THE GKEAT ERUPTION OF SAKUKA-JIMA IN 1914. 25 holociystalline rock was seen, probably making dykes through the basal vitroandesite (trachyandcsite) C. It contains abundantly raedio-phenocrysts (4 mm.) of hypersthene and sporadic niino- phenocrysts of olivine, which we frequently fail to find in many slides. The rock closely resembles the trachybasalt A^ of Fig. 3. The same rock makes up the flat- top of Sakkabira at the defunct Seto strait, and recurs at Karasu-jima. PI. XVI. Fig. 4. Besides the abundancy of phenocrystic hypersthene, which is rare in the writer's experience, the hypersthene is intergrown with peripheral and lamellar (100) augite. The light- colored trachyandesite C and the dark trachybasalt -li occur in close association, although the latter is decidedly of later origin. Both being widely distributed, it is an important pétrographie problem in southern Kyushii to make their relation clear. In analogy with the neighboring plateau-land, the rock which builds up the islet is of early Diluvial or late Tertiary age. ^oSm " '^) Hakama-goshi. (PI. III. Fig. 1 ; PI. XII. Fig. 3)— The low rectangular table-hill of Hakama-goshi or Shiro-yama in Geologic Map is of the size, and the same geologic formation and structure as the preceding. It was originally an island rising fi'om the bottom 20 fathoms deep, although it now forms a western liortion of Sakura-Jlma and a landing place from the city by the accumulation of talus and outwash from that mountain-slope, which filled up the space between this islet and the former Sakura-jima. Small crafts are said to have plied here even in historic times. It was occupied by the village of Koike, now entirely de- vastated and partially overflooded l:»y the deluge of the recent lava-flow. The lava-flow rushed down the western slope of Sakura-jima from the Yuno-hira vent to the sea, floodino; over the villaç^es 26 AET. 3. — B. KOTu ; of Yokoyama and Akobara, and touching the southern foot of Hakama-goshi on its forward movement. See Geologic Map, and PI. XII. Fig. 3. It is 72 m. high and consequently 30 m. lower than the city- castle, and rises from the sea-bottom 20 fathoms deep. It differs geologically from the island just described by the absence of the basal effusive G (p. 14) ; but the lower two- thirds are represented by the horizon B which is here barren in fossils, and built up Hkewise of sand mixed up with ash and pumice of the old hornblende -bearing hypersthene trachyandesite C. The upper one- third of the profile is again composed of lapilli beds A.^^ The whole complex dips slightly southeast. The islet is, like Oki-kojima, a fragment of plateau -land. K^su- n^ rpjjj^ LAVA-DEOWNED KaEASU-JIMA.'^ TllC Small pointed island (22 m.), now drowned and hidden under the recent lava, lay to the south of Hakama-goshi, and from the geographic position it is probable that it was a relic of the Plateau Formation like the latter. S. Kanat, who frequently 1) There is a shell bed of doubtful stratigmi^hic position. It may be a member of the lapilli bed A, or simply scattered over the top or slo^je of the hill. It is not yet decided whether the find is a kitchen middens, or deposited in the sea prior to the upheaval of the island. The species, all living, were determined by T. Iwakawa, of the Female Normal School of Tokyo, and they are as follows : 1. Fasciolaria trapezium Linn. 2. Strombus succinctus septimns Dnclos. 3. Strombus luhuanus Linn. 4. Conns quercinus Hwass. 5. Conus tes.sellatus Bom. ß. Conus miles Linn. (?. 7. Area decussata Sowb. 8. Chama imbricata Brod. i). Chama semipurpurata Lisch. (?) 10. Cypraea sp. 11. Turbo (Macrostoma) coronatus Omel. 12. Mactra sulcotaria Desh. 13. Septifer nicobaricus Chemn. 14. Vola laqueta Sowb. 15. Ostrea gigas (?) Thtmb. IG. Chlorostoma rugatum (Jould. 17. Gyrinium granifera Lam. 18. Phos senticosus Linn. 19. Columbella versicolor Sowb. 20. Nerita alficella Linn. 21. I'ectunculus vestitiis Wer. 22. Trochiis obeliscus G mol. I'lant remains are also said to be found in Hakama-goshi. 2) Karasu-jima or the ' Crow Island,' named jirobably fi'om the blaclc color of the rock of the island. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 27 landed on the spot prior to the recent eruption, informed the writer that the islet is built up of trachybasciH (PI. XVI. Fig. 4), as is the southerly-lying Oki-kojima, representing an effusive (Ai in Fig. 3) of the Plateau Formation. The writer received later a few chii)s of an odd specimen from the buried islet through the kindness of Mr. M. Uykda, who collected it many years ago. It is a dark-gvay crystalline trachybasalt of exactly the same type as that found in Oki-kojima, representing a certain horizon of the Plateau Formation, as the writer expected it to be (see page 25). It is incredible from a geological standpoint, that the islet welled out in one of the years 1475, 1476 or ? 1750, as recorded in some old documents, which were accepted with blind faith by nearly all the recent writers on Sakura-jima, nor is it by any means a parasitic cone, as some assume it to be {see p. 42.) kan-zé 4) Kanze.^' — A low sandy hoolc (PL III. Fig. 1 ; PI. XII. Fig. 3) in the channel between the city of Kagoshima and Sakura-jima, and lying to the south by southwest of the lava -drowned Karasu-jima. No solid rocks are exposed, and the origin of the islet remains unexplained. Float rocks scattered about are all brought from the neighborhood of the city. Lately, K. Ya:maguchi informed the writer that a solid rock, which probably makes the foundation, was brought up from the sea bottom of the neighborhood, and the specimen is black in color, resembling that which was found in Karasu-jima. PI. XVI. Fig. 4. From the proximity to the last-named and the similarity of the lithologie composition, Kanzé is in all probability of the same type and origin as Karasu-jima. K(S¥bu^ ^) Ko-JiMA OF KoKUBu. — A Small group (Figs, ö-ij) of islands, off the railway station of Kokubu, is composed of two 1) m m 28 APiT. -B. KOTO : sharply-pointed wooded islets on the north (Heda-Kojima/M24??z.) and south (Oki-Kojiraa, 100 ?».), both being built up of columnar ash-stone or piperno. A low patch (Beuten,-^ 40 m.) between tlie two is of the same jointed rock, silicified and pyritiferous, and colored brown through submarine decomposition, besides a cliff- rock (Ippai-jima) near the north island. The surround- ing sea is 20 fathoms deep. The whole group, therefore, consists of Pig. 5.- ^\'este^u view of Kukubn-Jvujiina. H.— Iltda-Kojiina. B. — Benten-jima. O.— Oki-Kojimn,. Fig. 6.- -Eastern view ol: Kolailm-Kojimn. 1) i^ H /> a 2) K. Yamaguchi {Geogr. Jour. Tokyo, No. 2, 1915) lately visited the islets and found a patch of tufl&te which is said to contain various diatoms. The same species recur in the Tertiary bed on the pass, east of the village of KajDd on the opposite coast of the mainland. Diatoms and coral beds are also said to occur in the above-mentioned Héda-Kojima, but their 'stratigraphie positions are entirely unlcnown. THE GREAT ERUrTION OF SAKURA-JIMA IN 1914. 29 basal trachyandesito C (Fv^. 3). PL XVI. Figs. 1, 2, and 3. The gray perlitic ß typo is to be seen in the south island. The upper, pyroclastic beds ^l and B of the Kagoshima type (Fig. 3) are not represented here. The same basal effusi^'e recurs along the coast (Nagahama) at the north of Kajiki on tlie land, whence the view was taken (Fig. ">), and this group is simply a detached mass of the basal portion of the plateau-land of the mainland. Since the last eruption, vulcanological literature has rapidly grown. An old book ' Zoku-Nihongi ' says : " In the 8th year of the Tembiù-Hùji era (764 a.d.) an eruption occurred at the frontier of Osumi and Satsuma. In the sea off Shinné village (the present Shikiué ?), three islands were formed of sand, done and lava.'" In another book these are apparently referred to the present group. The identification of the above historic record with the islands in question^^ was fully indorsed by recent writers on Sakura-jima. That this is, however, an erroneous interpretation will be clear by the statement made from a geological point, that the islands are simply detached blocks of an effusive Horizon G of the early Diluvial or late Tertiary age. [See p. 41). In saying this, the writer does not deny the truth of the said record, yet he could not locate the point of that eruption in Kagoshima Bay. In short, the islands of Oki-Kojima, Hakama-goshi, Karasu- jima, Kanzé and Ko-jima of Kokubu, dotting Kagoshima Ba3% are not only the same in age, in rocks and in structure among them- selves, but also exactly like those of the plateau land that encloses the bay (PI. IV. Fig. 1 and Text-fig. G). As has already been remarked, they are the wedged-up blocks of shattered bottom of 1) In the 'Explanatory Text to Sheet Kagoshima,' i>. 53, it is said that the island is ])robably parts of a ruined crater. Nishio speaks of them as the foundation-stone of the later Sakura-jima. J'etermanns MiüelL, 1014, S. 132 30 AlîT. 3. — B. KOTO : the bay. They are buoyed up in different degrees or, in other words, depressed in altitude with reference to the neighboring plateau after the manner of the formation of a plateau. Volcanoes are said to appear in diffuse manner in such a geologic pattern of great magnitude (p. 20). YÉNO-SHIMA 6) Yeno-shdia^^ (PL XII. Fig. 1, and Geologic Map). — In passing, it is to be noted that the islet Yéno-shima, lying 3 kiu. to the south of the defunct strait of Seto, is built up of uniformly gray compact trachyandesite.'^ PI. XVI. Fig. 1. It is simply a detached erosion remnant of the basal effusive of the neighboring plateau formation, and in no way connected with any lavas of Sakura-jima. SüBCETJsT YiQ j^Q^ a]3]3roach the question as to the foundation OF Sakuka- ^ ^ ^ •' "^^ of Sakura-jima. From what is said in the foregoing, the basement bed seems in all probability the same complex that makes up the four block-islands (1—3), and also the one that builds up the plateau-land of southern Kyûshû. As Sakura-jima is an island volcano, and is still young and not yet dissected, no opportunities are yet given to get insight into the inner structure of the volcano, and we have simply to conjecture the foundation in the realm of imagination. A unique specimen of recent éjecta was given the writer by Assistant Professor Kanai, of Kagoshima. It is a fritted granite.'^^ As granite -batholiths intruding the Mesozoic slate for- mation appear in Prov. Osumi, both may occur underneath the foundation ; but no slate has hitherto been discovered among recent éjecta except small fragments in trass.^^ The members of 1) ir / â 2) It is composed of plagioflase, orthoelase and hyperstliene in a crystalline gronndmass of ragged prismoids of angite, plagioclase-latlis and magnetite crystals, ■\\ith trace of globulitic and sanidinc -cement. 3) See ' Ejected Blocks of Biotite-graiiite " in Petrograpliical Part. See VI. XXIII. Figs. 1 and 3. THE GKEAT ERUPTION OF SAKURA-JIMA IN 1916. 31 the Plateau Formation (Fig. 3), which are supposed to make the foundation are not actually observed. The calcareous trass bombs among the éjecta, coated with black rind, may represent sandy tuffite B?^ From the cores of many stumped mountains it is ascertained that the depth of the local reservoir or lava-macula may be estimat- ed at 1,000 )iiP below the vents, and consequently, the thickness of the direct volcanic foundation may be assumed to be 1,000 ??i. Both the YuNO-HiEA and Xabe-yama vents of the recent eruption are located at 300 m. above sea-level, and the bottom of Kago- shima Bay is 100 fathoms or approximately 200 m. deep. fSetting apart 250-400 m. for the thickness of the Plateau Formation (Fig. 3), there still remains about 100 ??2. to make 1,000 vy^. for the subterranean location of the Mesozoic slatc'^'' formation, which is, as already stated, often intruded by granite (Fig. 2). It follows from the above that the lava reservoir of Sakura- jima probably lies within the terrane of the Mesozoic formation, for which the above-mentioned granitic projectile may be taken as a proof, and 500 m. below the bottom of the bay. Part II. The Volcano of Sakura-jima.'^ § I. Morphogeography. A) General.— The volcanic island of Sakura-jima is situated 1) See 'Ejecta of Trass' in retrographical Fart. 2) F. V. Wolff, ' Der Vulkanismus.' IM I., S. 331. 3) After a comparison of average compositions of igneous rocks and slate or shale, W.H. Hobbs arrived at the conclusion that lava, especially the Pacific type, originates through fusion of slate, the most easily-fusilde of sedimentaries. It is to be remarked that the lava- macula of Sakura-jima is probably located in the slate formation. ' Some Considerations con- cerning the Place and Origin of ^Macul;«.' Beiträge zur Geophysik, XU. Bd., l.)13, S. 330. 4) a. K. Nakashima, 'Explanatory Text to Sheet Kagoshima.' Geol. Surv. Japan, 1897. 82 ART. 3. — B, KOTO : within the narrow rift bay of Kagosliima, opposite and east of the city of the same name, and separated from it by water 3.7 km. wide and more than 25 fathoms deep (the maximum being 38). On the southeast side of the island there was, before the recent eruption, the still narrower channel (500 m.) of Seto, 30-40 i'athoms deep, where during the explosion of 1779 a floating sheet of pumice, 5 feet thick, covered the whole stretch, so that two A te^rror-stricken islanders fled on foot to the opj^osite coast of Osumi, while the third, following the two pioneers, sunk to the bottom and was seen no more. This channel was completely choked up by the lava-flows on February 1st (some say on the 29th January), 1914. The insular composite konide of Sakura-jima rises from a depth of 100 fathoms, having the equatorial and meridional diameters of 9.9 and 8.0 km. respectively.'^ It has a circumference of 38.7 km., with an area of 75.6 sq. km., and a height of about 1,000 ?7i.,'' with graceful piano slope below 1000 7?z. all round, b. I. Friedlaender. ' Ueber einige jupanisclie "S'iükane.' Miftell. d. J). Gesell, für Natur- uml Völkerkunde Ostasiens, Bd. XXII. Teil 2, S. 103, 1910. The author i>resents briefly the views aud observations expressed by S. Watanabé. c. Yamasald-Sato, ' Geography of Japan,' Vol. VIII. p. 223, Tokyo, 1!)09. (Japanese) d. Uyéda, Salmra-jima before and after the Eruption of 1914.' Jour. Geoijr. Soc. Tokyo, 1914, p. 431 et seq. (Japanese) 1) These numbers refer to date prior to the eruijtion of Jan. 12th, 1914. 2) The area and volume calculated by stud. E. Oda after Simpson's formula : a. Area 75.59 sq. km. (71.90 „ „ after D. Sato) h. Height traced on Contour-line. Volume above Sea-level. 0 to 200 TJi 9.04 cub. km. 200 „ 400 3.863 400 „ 600 1.683 600 „ 800 0.937 800 „ 1,000 0.390 1,000 „ top 0.036 Total 15.949 cub. km. It is nearly the same height as the volcano of Kaimon. The volume is one-si.'ity- fourth of Tuji-san, the latter Vicing 1,025 cub. km. F. Ômori gives the figures 26.5 cub. km. for the volume of Sakura-jima. THE (iKEAT ERUFriOX OF SAKURA-.TIMV IN 1914. 33 excepting tlie western side ; but seen from the north and south tlie mountain appears perfectly conical. See Text-fig. 10 a. The coastal ilat of talus and wash plain supports 11) village groups of 0,134 ménages, sustaining a farming and fishery popu- lation of 27,116. The sunny island produces oranges and gigantic radishes of half a metre in diameter. Horses, pigs and domestic fowls are raised on a large scale.'-* B) The Building-up of the Volcano and Topography.^ Sakura-jima is apparently a simple konide sloping at 5° to 30"^, and truncated on the top with a serrated wall (PI. II. Figs. 1-3 ; PL III. Figs. 1-2) ; but the apical crater is really a triple one arranged in meridional direction. The volcano is also, geologically 1) 2'Ae vegetable zone of ^ft. Sakura-jima. —Fvom. a list of plants registered by the teaching stafiE of the Higher ('ollege of nemlrology, Kagoshima, vre see there are ncj M-ell-delined vegetable zones on the konide of Saktira-jinia as there are on many other moiintains, yet ionr zones can he recognized according to the predomin.ant elements See Fig. 7 on page 34. I. The apical zone embracing the belt from the 'JOO-metre contonr-line upward to the top of 1,069 m. Among the scanty vegetation in the crateral area may l« mentioned rashes and brush- ■vvood — Imperatn nrunäinacen, Cyr. (chigaki), Misc/mthus sinensis, Ander.s. (suziiki), associated with Fanicum samjubuile, L. (me-hijiwa) and Eleuslne curacana, Gaertn. (shikoloi-bie). On the east slope of the middle crater there is a localized forest of Alnus fiitna, SZ. (yasha-bushi). II. The next one is the 500 to i)00-metre zone, represented by a mixed forest of the preceding and the next follo'wlng. Principal elements are Hydrangea virens, Sieb, (galai-utsugi), H. paniculata, Sieb, (nori-utsngi), DieriMla floribunda, SZ. (nishi-utsugi). Viburnum dilatatum, Th. (gama-zumi), Ilex crenata, Th. (inu-tsugé), and Salix saidtana. Seem, (yama-yanagi). in. The zone of Finns Thunhergii, Pari, (kiiro-matsu), partly natural and partly i^lanted, Teaches the 500-metre contour above, wliile on the east it is alrea^ly delimited at the 300-metre contour, due probably to its being on the lee of the prevailing winds which cause drought on this side. The lower boundary of the zone lies between 100 to 150 metres. The belt is also a mixed forest, associating with Cryptoineria japonica, Don. (sugi), Ligustrum japonicum, Th. (nezumi-mochi), Thea japonica, (L) î^ois. (tsubald), Stauntoni'i hexaphylla. Dene. (nubé). IV. The remainder of the area is the narrow coastal zone on the west and north and a small strip on the south, nearly all being under cultivation. The gardens and orchards are laid out in coarse volcanic soil on slopes. The staple i^roducts are Citrullus vulgaris, Schrad. (suika), Saccharum officinarwn, L. (sato-kibi), Ipomaea batatas, Poir. var. edulis, Mak. (satsuma-imo), JRaphanus sntivus, L. (the well-known large radish), Eriobotrya japonica, Lindl. (biwa), Frunus communis, Huds. (sumomo), and a number of oranges, viz.. Citrus aurantium, L. var. tacMbana, Mak.; var. japonica. Hook (kinkan) ; subsp. sinensis, Engle. (natsu-mikan) ; subsp. nohilis, ^lak. (mikan), and a .Japanese persimmon Diospjyros kaki, L. 34 AKT. 3. — B, KOTO : speaking, a single gigantic konide, having an altitude of 1,133.5 ???., and standing on an irregularly round base of 8 Lm. in diameter, which approximately coincides with the present dimension of the island. The main crater-cone is Kita-dakk {MUiacJii) or the north crater-cone (homate). rr^ill^fN PROV, OSmil 'Vt, Û man Z,n, of .\L,.Jt:„.:i U,^ll ,.,i (Eau // C.^1.1 Zm unJ^ Cu/uriù^ - I 1 V V,/hg,:, »„J O^UJ ) Fig. 1. — Sketcli-map illustrating vegetable zones. Later, the two other apical craters, the south and tlie middle, were created, the former being comparable in magnitude witli the main ; and the total morphological aspect of tlx^ mountain top presents that of a triple volcano, although the two later ones are really parasitic to the main, the iiorth cone. The reason why they are so ca.n be readily understood from the wide distribution {sec Geologic Map) of the Hght-brownish (jvmj lava {Mihachl lava), which characterizes the material of the main body of the volcano and the crater- wall of the north cone. THE GIÎEAT ERUPTION OF SAKÜRA-JIMA IN 1914. 35 The fresh South Cone which rests upon the shoulder of the trunk, poured forth J)J(tck lava {Mmaini-dalié lava) sheeting the whole southern flank down to the shore. The Middle Cone, on the other hand, also sent down black lava (Fiitannka lava) only to the (?) eastern slope. In short, the later two cones are adventive to the northern main. In the following the writer will shortly characterize the three craters above-mentioned , KiTA-DAKK a) The northern coue, Kita-daké or Miliachi (PL III. Fig. 2), has a rather deep bottom with a diameter of 500 to 600 m. (iuuer diameters 212-200 ;y^) and a depth of 100 m. (80 m.), and the precipitous wall 1,133 m. high which is the highest point of the volcanic island (PL IL Figs. 1-2). The pumiceons bottom is now clothed with mosses ; and benches of la^'a are exposed to view on the inner cliftj which is a little ruptmed on the east and west sides. Naka-daké h) A little apart, southwards from the preceding, lies the shallow Naka-dahé (Futanaka or Futatsu-ai) or the middle pit (PL III. Fig. 2 ; PL Y. Fig. 1, M) with a N.-S. diameter of 400 m., an E.-W. of 200 m., and a depth t)f only 35 m., the highest point of the wall being 1,000 m. It sits npon the cliiT of the crater-bottom, 250 m. deep, of the South Cone or Minami-daké. The ash-covered bottom is clothed with mosses and shrul)s. In ancient times it -was full of water, and it was said to rise and fall ^\ii\\ the tides. That it is comparatively a neiu one is testified by the x^resence of vents of solfataras, although they are now entirely extinct. The eruption of 1779 started in this veuthole, or somewhere from its outer slope. ^^ Watanabe-^ says, this is tlie oldest of the three pits, ^\'hile the late 1) It was commonly said at the time tliat the eruption began at Futanaka (^ tfi) or the crotch between the north and soath craters. Some say at the southern one. We need not ha very particxilar about the location of the first subaërial eruption; for, at that time the entire apical region or indeed the whole mountain was in a state of paroxysmal conviilsion. 2) S. Watanabé, ' On the Volcano of Sakura-jima.' Jour. Geoijr. Tokyo. Ko. 289, 1913. (Japanese). Friedlaender is of the same view as Watanalie. 36 ART, 3. — 13. KOTÙ : Nakashiina^^ (Le.) Avas inclined t(j believe this one to be younger tlian Ivita-dake. The Avriter consider.? the north pit to be the first-born, then comes the southern, and the middle pit to l)e the yoinu/esf of the Fig. 8.- — Sllo^^ing the relation of the North, Middle and South Cones, \aewed from the "western slope. Photo, by Yaniagixchi. three, as we can see the middle pit 31 re.sting directly upon tlie shoulder of the southern pit craters on the cirque-like explosion ('?) cliff near the top on the west, as in the above figure 8. DAKÉ ' ') ^^^® southernmost, elliptical Minami-dcikc (l,0G9-6 w.) is the lowest and the active one (PI. m. Fig. 2 ; PI. V. Fig. 1). Form- erly it was called ' Shiro-mizu ' or white water, l)ut now peoj^le call it Moye-d-ih} or 'burning pit,' and it is the Lirgest crater, having a 1) The writer fully endcjrsos the \ie\v expressed l)y the Lite Nakashima from a i^etrographic standpoint, since the light C(jlored salic hyi^ersthene-andesite of the n(jrth cone greatly differs from the black femic sempatic two-pyroxene andésites of the middle and south cones, which sometimes even contain an accessory of olivine. The only question left for future consideration is : AVliich has the precedence, the midly a ])Ortion of the stemmed-up lava-flmv of the south cone. THE GÜEAT EÜüPTION Ol' SAKUEA-JIMA IN 1914. 37 diiimeter üf 650 in. by 450 m. tiiul n depth of 250 m.,^^ eucireled with SI sharp ridge of pitch-black andésite, sometimes rusty browu. It is lx)imded inside with precipitous chffs, so that no one can reach the bottom, where twenty or thiiiy years ago a pond existed which is noAV entirely dry. On the southwest and north chffs, built up of mud breccia mixed with pumice, capped with tabular or columnar solid lava, sulphurous fumes issue but not violently. Gulch About 500 m. landward from the shore near Yuno- hama at the southern foot of Minami-daké, there is a spot where steam and carbonic acid mixed with sulphur dioxide issue from tlie bottom of a dry ravine. Insects and birds and even men were asphyxiated on approaching the aperture of the mofette. The place is called by the people Dokudani or ' poison gulch.' The form of the narrow lava-conduit of whole Sakura-jima PoLYAxiAL js what F. V. Wolft' styles the ' polyaxial central type.' Type rpj^^ ^^^^^ apical advcutive pit-craters, the middle and southern, are only accidentally formed Ijy l)ranchings of the principal canal, as may be surmised by an inspection of the annexed Geologic Map. The linear compound volcano so formed has attained roughly a maximal height of 1,000 ??z., which is, tJie icriter believes, the average height of modem volcanoes in Japan, corresponding to the size of the konide, the nature of the magma of intermediate acidity, and the amount of energy of the magma reservoir which is usually assumed to lie under 1,000?». from the ventholes. Topo- TopoGEAPHic FEATURES. — From the mode of for- GKAPHIC Featuke mation as well as from the material which built up Sakura-jima, we are able to recognize the topographic form of 1) The maximum depth of the crater-bottom of other andesite-Tolciinoes of Jaimu seems to the writer to be 250 to 300 w., without the danSer of disruption of the boily of a volcano. 38 AKT. 3. — B. KOTJ the mowita in-top on maps, or from the view either from west or east. From the city of Kagoshima (PI. III. Fig. 2) we see a deep gulch beyond the round top (553 m.) of Hikino-hira. "^c^N™^ Parasitic cones will be briefly sketched elsewhere.'^ Here the southern flank of the main (north) body was ruptured Fig. 9. — Eastern view of the three cones. before tlie building up of the south cone. The two possess topographically independent forms. A view from the east (Fig. 0) likewise presents a distinct line of demarkation between the main, north cone N and the middle parasitic one M. After the piling up of the three apical cones, whereby the volcano attained its maximal height and perfection, new vents were opened a little below and outside the apical craters. They are the 11 a n k eruptions, when lavas poin-ed down in pericentric flows {see Geologic Map). The time of their activity falls mostly in the historic period. The volcano of Sakura-jima has entered since then into the destructive phase and senile age, and its height 1) The parasitic cones of Nabé-yama and Yébino-tsiika on the east coast will be briefly toiiched under the heading : - The Lavas of the lîummei Eruption.' The problematic i)urasitic cone of Hildno-hira (Hyoku-oka), having the appearance of a' monadnock. will be spoken of in connection with the 'lavas of Kita-daké.' THi: GREAT ERLTTION OF SAKUKA-JBIA IN 1914. 39 ^vill ill futures gradually decrease, "^rhrougli repeated eruptions, which reuiarkably avoided the area (jf earlier flows, the former graceful form hecame irregular, the structure complicate, and the moimt grew and will grow sidewards. On the other hand, the relief and plan are in course of time l)eing modified and simplified l)y subaëriju a,nd ])luvial talus formations around the fringes of the volcanic island, especially on the north and west coasts. The topography of the mountain s^ope or piano is therefore entirely governed by pericentric radial flows. They are geologically recent rocks not yet disintegrated, all l)eing black slaggy block- lavas which can scarcely afford strong foothold to trees and shrubs. Only the inter-lavaless strips and the talus formation as well as the outwash plain near shore are available land for pasturage and horticulture, for which the island is tolerably well known (Text-fig. 7, Zones IV. and Y.). Arable land is reduced to a minimum, owing to the poverty of soil and the scarcity of water. C) Hydrography.— From the geologic structure and the topographic form as outlined above, we can w^ell understand, that only the radial gullies on the slope serve as draining channels for meteoric water. On the west and north talus and torrential streams spread and deposit gravelly alluvial fans. As the ground is full of clefts, or of loose nature, the w^ater is readily soaked in within a short distance. Consequently, there exist no permanent spkiîjgs streams on the island. Copious sprln(/s,^^ however, issue at some places on shore from the under gravel bed, or one can get plenty of ground water in shallow wells dug at the lower end of wash plains. 1) See ' Changes in the efflux of hot-springs,' p. 17, and ' Hot -springs," pp. 56, 57, itud lastly, 'Hot and cold s])rings,' ]>. 62. 40 ART. 3. — lî. KOTü : Occasionally cloudbursts precipitate unusual rainfall, which runs down from the steep mountain slope in torrents, excavating narrow and deep channels and carrying all that is found on the way. The havoc brought about by high water frequently causes the people great distress. The calamity incurred from tliis cause after a great eruption is very heavy, owing to the suspension in the w^ater of the hght surface deposits of ash and pumice which were thrown out and spread about from the ventholes. To cite an example, there occurred on November 20th, 1779, a great rush of mud flood on the north coast, which devastated the villages of Matsura and Futamata. It was two weeks after, but not at the time of, the great eruption of the An-ei era. Mud flood Pcoplc Call the mud flood 1/ama-skhco,^^ or ' mountain- (ïAMA- sHiwo) tide,' which is always a inore dreadful calamity to the inhabitants than the terror of the actual eruption. Four deep dry ravines on the northern slope, well shown on the Geologic Map, are the indelible scars left by the excavation of that tremendous non-volcanic yama-shiiro. They are the Matsura- gawara and Futa- mata-gawara. There are many others not less notorious, on the northwest and southwest coasts, and always in densely populated quarters. The same unhappy events happened after the recent eruption even outside of the island, causing deluges of mud, and changing the course of streams. Melioration-works with timber-bars, hurdle or tress-works, brushwood bundles or gabions are not used w^ith the exception of dams of piled stones, as the torrent directly empties itself into the surrounding seas without inflicting much devastation to the low lying tract, which is here limited only to a small area on shore. The geography of Sakura-jima is lacking in names for rivers, 1) \IL\ \^ It is the Jökellü'i of leelanil. Iddiiigs, • The Problems of Volcixnism,' p. 6. THE GüEAT uliUPTIuN OF SAKURA-JIMA IN 1914. 41 1ÎIVEKI.ESS i^^j^ instead tho term ' kawabara ' (spoken lunnthai'ii) is of frequent occurrence, which signifies valley train or wash plain ; for, in ordinary days not a drop of water is found in the percolating river-bottoms, e.g., the ' kawabara ' of Take, Akamizu and Nojiri, the latter two, however, are now overfloodcd by the recent lava- sheet from Yuno-liira on the west side. §11. The Eruptions at Sakura-jima prior to 1914. Eeuptiox- PEr.IOD I. A) The Eruption-periods I.-I V.— Tho building up of Sakura- jima is stated in various works to have taken place in 71^ the years 708 a.d. (1st year of Wadô era), 716 (2nd 717 718 year of Eeiki era), 717 (1st year of Yôrô era), and 718. The island is said to have welled up and risen in a night.^^ Eruption- ' ^\^\^q frontier of Osumi and Satsuîiia was in the PEKIOD II. ^*^^ 8th year of the Tembiô-Hôji era enveloped in dark clouds in which the lightning darted about. After seven days it became clear. In tlio sea of Shinni-mura (present Shikiné ?) three islands were formed of sand, stone and lava, the eruption being accompanied by flames. The islands so formed appear conjointly like an open shed. Sixty-two divisions of houses were destroyed and eighty people killed ' (Zoku-Nihoiigi). Other books say that the islands referred to are Kamizukuri or Kojima, by Avhicli is meant the ruin-like Kojima group off Kokubu Station. In July-August, 766, they were visited by terrible earthquakes. That the whole story was an erroneous interpretation of ancient records is elucidated elsewhere (p. 29). 1) From 708 to 718 Ivjigosliinui Bay seems (o Lave been the seat of violent volcanic activity, constitutiug one eruptive period, having in common a hypoceutre of acti^'ity. It is impossible for the writer to locate the exact ]ioints of vents, nor can he say whether they lie excliLsively within the area of the jiresent Saknra jima or not. 42 ART. B. — B. KOTü : ebuption- ]7^'^Y a l(~)nf!: interval of seven centuries Sakura-iima PERIOD III. - O J BujmEi remained apparently in (inietude. The third period of ^iS''^^ activity began in 1408 (2nd year of Onin era) with a moderate explosion on the soutliern top-crater. Another eruption 1471 occurred on October 2r)th, 1471 (3rd year of Bummei era) near Kurokami on the east coast, ejecting stones, sands, and overflooding lavas, resulting in the formation of Cape Omoyé-zaki.^' 1475 On September loth, 1475 (7tli year of Bummei) a violent eruption took place above Nojiri on the southwest coast, in- flicting great damage to the inhabitants and domestic cattle. Cape Moyé-zaki, which lies between Nojiri and Yuno,'^ was then formed. Ash fell during five days, and Karasu-jima rose from the sea.''^ 1476 An ash explosion of the same intensity occurred in 1470 (8th year of the same era). On September 29th it culminated in the greatest outburst. During five days before the event the land was incessantly shaken by weak and strong earthquakes. Even- tually the top (which one ?) was in flames and disrupted, throwing out rains of burning stones and ejecting ash and pumice for seven days, which envelqped Kagoshima and the neighboring districts, turning day into night. It was followed on October 6th by lava- flows, which caused an increase of land 8 km. in circumference on the southwest coast (Nojiri-Furusato and (?) adjoining land). Many l)eople and animals were kifled and wounded. Afterwards Oko- shima'^ rose from the sea off the shore of Yuno. The eiuptioa of U7U seems to the icriter to have been of the same type as tliat of the An-el era, shortly to be described, thougli it was on a somewhat smaller scale t!ian the latter. 1) It is the lava field of Urano-mayé. See (icologio Map. 2) The statements in old documents concerning the activity of 1475 and 1476 seem some- what redundant and rei^etitive, and the succession of events is not entirely clear to the writer. 3) Geologically speaking, Karasu-jima is of Diluvial age (p. 20). 4) This traditional recjrd is entirely unfouiided (i). 23). THE GREAT EKLTTION OF .SAKIJllA-JIMA IN 1914. 43 1478 The after-eiTects continued till 1478, when the moun- tain erupted again and the slu^re of Fukuyama on the op}30.site coast of Osunii was converted into a desert by rains of ash and pumice. During the whole period the Kirishima group, tlie northern neighbor, kept a profound silence, or at least there is no record of its activity. In loo 4 (23i'd 3^ear of Tembun era) and 1556 (2nd year of Koji era), Kirishima was, however, active. In 1615 (1st year of Genua era) Kaimon, this time the southern neighbor, made a great eruption. Thus, while Kirishima and Kaimon were displaying activity from 1554 to 1615, Sakura- jima was not in the least affected. Eruption- Qu April Ist, l(i42 (19th year of Kwan-ei era), an PEEIOD Iv. i \ J ; ^'^'^^ eruption took place. 1678 On February 29th, 1678 (6th year of Em-pô era), an eruption took place. i"06 In January, 1706'^ (2nd year of Hô-ei era), an erup- tion (?) took place. 1742 In April, 1742 (2nd year of Kwan-pô era), an erup- tion took place. 17*9 Li Septemhcr, 1749 (2nd year of Kwan-yen era), . ML Oliiva-ycima^^ erupted xiolently, spreadlncj an apron of lara-ßow just to the south of Ilik'nio-hira (Hiyoku-oka). 1756 On September 9th, 1756 (6th year of Hô-réki era), a hot-spring in Yolwyama gushed out, but has since disappeared. 1766 On May 21st, 1766 (3rd year of Mei-wa era), a great inundation by mud-flows took place from, the top (which crater ?). 1) In reatling the phrase, *■ the mount ' bnrnt ' or erupted," we must guard agiinst falling into mistake. To cite an exam^jle, the writer found in the record of the same year the following clause : The Sakura-jima shrine located on the sloi)e of the south crater about nine- tenths from shore was burnt by accident from the tobacco pipe of a farmer of Akamizu. 2) The extent of this Ohira lava-field is cartographically represented in Geologic ^lap. 44 ART. 3. — B. KOTO Eeuptiok- PEKIOD V. (1779-179!)) ^ c B) The Eruption-period V. — a) The Eruption of the An-ei Eea. The ac- tivity of the An-ei ei'iV^ was the most terrible one ever recorded of Sakura-jima, especially from 1779'^ to 1781, though it was persistently turbulent for twenty years, gradually declining in in- tensity and finally coming to rest in 1799, having t^xhausted its volcanic energy. 1779 From Novem- \)QV 7th, 1779, Kagoshima und the environs in a radius of about 40 Im. were violently shaken, and people became very much excited with fear. From 10 \ W to noon of fho ^iQ- iO •"'■ — Southern view of the first iihase of eriiption fit noon, on the 8th November, 1779, as seen from next day (8th) the water Tanimi, Prov. Ôsuml, the active center being at this moment the south crater-cone. Copy of a sketch by m all the wells of Sakura- s. Mchi. k< U.rt-.( 1) There are several works, both Japanese and foreign (Milne, Friedlaender etc.) on the erui>tion of An-ei, all telling the same story. The present version is chieily leased on fhiri-sanko, H ' Compentlium of the Geography of Hyûga. Ôsumi, and Satsuma.' 1898, Kagoshima. A ftül list of the literature is given in the ' Explanatory Text to the Sheet Kagoshima,' by the late K. Nakashima, and in the late Milne's ' The Volcanoes of Japan,' Trans. Seis. Soc, IX. 1880. 2) From 1772 to 18S0 the Kirishima group remained remarkably tran(jnil. THE GKEAT EKUl'TION OF SAKÜ11A-JIMA IN 1914. 45 / jiina boiled up, spouting at several points, and the color of the sea became purple. Soon after 2 r.M. a thread of wliite cloud rose from tlie south crater, and then black smoke was thrown out in cauliflower form'^ from the dope between the middle and south top-craters, followed immediately by tremendous roaring and violent quaking. In the house one felt as if sitting on a rocking chair, and out of doors it was like being on the sea. When persons lay down they were rolled about, and when they stood up they were thrown down, and when they sat down they swung to and fro. Flames shot out from hoth (Figs. 10/>, 10c) sides of the mountain ejecting pumice and sending down avalanches of mud,'^ at first at Shirahama near the nor- thern coast (see Geologic Fig. lO b. — Sakura-jima in eruption in 1779, ^-iewed from Kagoshima. Four active centers, the southern- most (on the right) being on the slope at Anei-zau vent ; the second at the crater of the south cone ; the third and fourth (on the left) on the northea- stern slope of the north cone (of. Fig. lOc). The northern vents were very active on November 7, the southern on the next day. Copy of a sketch in posses.sion of ]\I. Hamasaki in Kagoshima. 1) According to A. Imamura, the height of the column of smoke-cloud was measured at that time by the compilers of an almanac in Kagoshima to be 1"2,029 m. (3 rl 2 c/to 16 fcen)— an altitude that is already in the windless and cloudless zone of constant temiaerature. 2) It seems highly probable that the ejection of mud- flow took place at the early phase of activity from one of the top-craters, either the middle or the south. 46 ART. 3. — B. K0T5 Map) and thon at (?) Minami- daira. Soon the island was en- veloped in dark smoke periodi- cally illuminated with flashes of lightning and projectiles of red-hot stones like meteors witli tails. This threa- tening state last- ed for five days, when the inten- sity abated somewhat, though for a month every 3 or 4 hours, or every 2 or 3 days, it burnt and roared. The wdiole aspect of the mountain island was then entirely changed, elevations were engulfed and hollows filled up w^ith ash (?) and later by lava-flows. On this calamitous occasion 153 people were killed and man\' more w^ounded ; 1,576 horses and 135 oxen w^ere also killed, besides other domestic cattle and fowls. Two thousand islanders fled to Kagoshima, and on the other liand the citizens were thrcnvn into confusion by an unfounded rumor of ' tsunami ' or encroaching boisterous sea waves. Heavy damage resulted in Yuno and Arimura on the south, Kurokami on the east, and Komcn on the northeast coast. Headlands of lac a wttc formed in Komen and Kurokami (see Text-fig. 11), both bearing the Fig. IQ .-. a — Lava ends of historic lavas prior to 177!J. V — Vents of the An-éi eruption, 1779. Sketch-map by S. Izichi at the time of the An-ei activity. THE Gr.EAÏ ERUPTION O]-' SAKUKA-JIMA IN l'J14. 47 saino na.iuo. (?) Movézaki or ' burning capo.' On the southern flank a. lava stroain crawled down to the shore from a large explosion-hollow of An-ei-Zan or Shin-moyk (see Geologic Map), forming Cape Tatsusaki. In Arimura a new lara-hland rose but disappeared later. Changes ^ hotsimin I ccascd to flow ill Arimura,^^ while IX THE -^ •' ^^hJ?- *^"^ two new ones began to gush forth on shore at the east and west ends of the lava-held between Furusato and Yuno-hama. I'hey originated probably from meteoric water descending the slope, being warmed on the way by cooling lava and later mixed with sea water. The temperature and level were said to have changed with the phases of tides. There is another hot-spring, 36 C, wdiich is located at the bottom of wdiat seems to have been once a small cove in Kuro- kami. The spring is of an old date and of saline nature with alkaline reaction. It once ceased to flow, but revived in 1771). This is also located at the margin of the An-ei flow on shore. Ash-fall Asli fell ill enomious quautitics, especially on the Osumi side of Kagoshima Bay, vu., the vfllages of Tarumi, Ushiné and Fukuyama, and a sheet of pumice covered the bay to a thickness of ö or 6 feet, so that the terror-stricken people fled on foot, on November Otli, to the opposite Osumi coast over the narrow channel of Seto, and wild boars crossed to the Kagoshima side (the high flat of Yoshino, sec Geologic Map) over the floating pumice. During the eruption 'the whole island of Kyûshû, tlie provinces of Isé, Shinia, Owari, Mikawa, and farther east became dark and ashes fell there. On the 17th November asli fell in 1) The hot-spring revived during 1835-'36. It is of saline nature witli a weak acidic reaction and little content of carbonic acid, the temperature being 39'C. The two new ones have a temperature of 45'C., and saline cliaraeter with alkaline reactions. See p. 56. 48 AKT. 'S. — B. KOTÙ : A Tokyo, u distance of 1,000 km. from the vent ; m Osaka on the Oth. Patches of pumice were even seen floating on the shore of Inaba, Hoki, Tajima, and Tamba in the Japan Sea. h) The Five New Islands.^^ In consequence of the eruption of the An-ei era, many islands made their appearance within a year. During the eruption Sakura-jima was entirely enveloped in dust smoke and cloud. When the air cleared the island came into sight, five days after the ever -memorable outburst of the 1779 8th of November, 1779. After a few days (November 11th or 12th) from the bottom'^ of the sea off the northeast shore of Sakura-jima imearthly sounds issued, like bellowing, accompanied with seething and broiling of the sea, ejecting of stones, pumice and mud (fine pumice splinters), resulting in the creation of shallow banks and small islands, which constantly changed their forms and positions. At first, fountains of yellowish mud rose to a height of 30 to 40 feet, which fell l^ack and flooded^^ villages on the shore of Sakura-jima and the adjoining lands of Kagoshima Bay. The islands so formed became finally fixed. They are the following (see Fig. 11) : iNOKo-jntA jVo. 1 Avas created ou November 21st, 1779, near the (Nos. 1 AND 2) I77'j shore, ouly 327 m. (3 cJ/o) from the coast of the village Komeu (Mnkomeii)/^ Its diameter was about 100 in., but height only 1) As mucli doubt M-as cast itpou the origin of the new-born islands, the writer in April, 1914, made a short visit to the islets and rocks in order to collect rock-specimens and also to study their structure. 2) Said to have been 70 fathoms deep in 1779, and this will by no means appear an exaggeration if .inyone examines modern charts. 3) People call this tsunami or shiwo-ogê (-i^JJ fg f] signifying sea-wave or tidal -waxe. 4) It is to be remarked that the distances and directions mentioned in Chiri-Sankô are not entirely intelligible to the reader v,ho takes up a modern map for reference. The writer thinks this is solely due to the fact that the once large -village of JMukômen along the coastal ])lain jierished under lava-flows. The buried part, as villagers told the writer, stretches between the l^resent Urano-mayc and Kômen, which space is now covered with (•()nfused masses of block- lava, as indicated in figure 11. THE GEEAT ERUPTION OF SAKURA-JIMA TN 3914. 49 3 m. This epheinerul Lipilli island disappeared on August 1st, 1780. No. 2 rose from the sea on November 22nd at a distance of 138 m. (1 cJto 16 hvt) eastwards from No. 1. This pumiceous-rocky mass (hyper- stheue-andesite) forms the present Inoko-JIMA^^ which after the late erup- tion appears above water only at low tide. The islet or pointed rock oft' Oseko-zaki in Text-fig. 12 is the one here referred to. A shoal, lying 200 77?. south of Inoko-jima, may represent the ephemeral island No. 1. ■■> f , , Yebisu^ima > \ ^^ «o "o .) ^ "'^vvo-jima / ,. — ■ / ,(Nos.«.^) Fig. 11. — An-ei lava-fieW find new islands to the northeast of Saknra-jima. 1) The writer identified the new islands on modern maps from circumstantial descriptions given in a book, ' Chiri-Sanl^o.' Piecently, K. Yamaguchi {Geogr. Jour\ Tokyo, Jan. 1915) went over the same line of study as the writer, with a result which does not harmonize well with the one at which the writer arrived. He wishes to leave the judgment to others who may care to pursue the same course of historic interpretation. . 50 ART. Ô. — B. KOTa: Fig 12. — The Inoko-jima rock in foreground and Silin- jima >S' in distance — the two new islets of 177Ü-1780. (After Yamaguchi.) Iwô-JiMA Nos. 3 and 4. Tlie former a,i>i^eai'ed on December (Nos. 3-4) 1779-1780 13th, 1779, at a distauce of 1,600 m. (15 c//o) eastward, and the latter near by (6 eJio) on Jaunaiy 15th, 1780 ; both are rocky, and once emitted Ja sul- phurous odor. The two afterwards coalesced into one. It is Iwo-JIMA (Kuro-shima or Se- jima) — the ' sulphur island.' This compound islet appears hook-shaped during low tide, as if it were a sub- merged rim of a small crater- wall. This is the only new islet built up of Avliite-spotted black rock'^ that the writer has seen among the new islets. See Fig. 13, I. Shin-jima 2^0. Ô. Two islands welled out on May 11th, 1780, in (No. 5) J ' ' 1780 the direction S.S.W. at a distauce of about 1,500 ?w. (14 <7. 14). A slightly sjiliue Avater seeps out fi'oin the jnnction of the two beds on the south shore, which induced five families from Akamizu of Sakura-jima to settle here twenty years after its formation. A ])umiceous gray lava-How covers Fig. 13.— Three new islands, Shin-jima, IwO-jimfi, luid \"('l)isn-jiiiia, nt 17v!i Jv.so. (After YiimiigTicbi.) l)oth edges of the central low stretch, which seems to the writer to be a ruined crater-bottom of a submarine vent. Tlie cliffs which bound the island are solely due to marine abrasion. On August 9, 1786, a storm swept away the south-eastern headland leaving a shallow wliich was once occupied by hind. 1) K. Yamagtjchi {loc. cit., footnote p. 49) is inclined to believe the island of Shin-jima to 1)6 of an old date in spite of frequent citations about its new creation during the An-ei erup- tions in the literature of that ]ieriod. Sandy tiiffite and lapilli lieds, which Yajiaguchi mentioned, build up a large portion of the iîslaud, and closely resemble the Diluvial Plateau Formation of Kagoshima ; but if these beds are really identical with the latter, then the island must be considered to be a detached geologic block of the Plateau Formation, similar to ()ki-Kojima and four other islands briefly outlined (p. 23-30). The writer, however, found no crystals of hornblende or biotite in shore sand in his collection, which are important constituents of the basal andésite from which the old tuffite and lapilli beds are derived elsewhere, iloreover, the pumiceous andésite, which intrudes and partially covers the underlying sedimentaries, is of the same Idnd as that of the An-ei lava — a fact indicating close genetic relation with the neighboring islets raised during the An-ei eruption. A remarkable fact discovered by Yamaguchi, is, however, the jiresence of moUuscan shells and diatoms entombed in the surface soil, 44 m. high, which, according to his interpretati(m, are not of recent origin. In short, the question as t(j the origin of Shin-jima awaits further researches. 52 AKT. 3. — B. KOTÔ : ,. , iV^o. U welled out on Jimo 12th, 1780, at a distance of (Nos. 6.7,8) 1^500 TO. (over 14 cAo) in the direction E. 15^ N. from No. 5. Nos. 7 and S. The fh'st rose on September 29th in the direction 8. 45°E. from No. 6, and the second (No. 8) near by on November 9th. Afterward Nos. 7 and 8 united, and later No. 6 also coalesced with them so that this compound island, resulting from amalgamation of the component islets Nos. 7, 8 and G, is designated by the people No. 6. This is Yebisu-jima (Nakano-shima), and is like Iwô-jima built up of pumiceous andésite^ ^ clods, thinly covered Avitli pine, and fringed with a shell-mixed sandy shore. This is the only islet that looks white from the distance. A slight depression in the centre may be an indication of a defimct vent (Y in Text-fig. 13). The islands Inoko-jima (Nos. 1-2), Iwô-jiraa (Nos. 3-4), Sliin-jima (No. 5), and Yebisu-jima (Nos. (î-S) are called the New Five Islands, tliongh in reality there are only four independent islets."^ Nn-BAMA In passing, the writer may mention the islet of Hama-shima""^ (Nii-shima) off Kurokami village. See Geologic Map. It is a mere lapilli bank, elliptical in form. It is said to have been formed by submarine eruption during the An-ei activity. The old lapilli accumulation is commingled with recent éjecta, in which the writer picked up pieces of cordierite-bearing ones 1) The rock is diül-grfiy wbite-i)ori)byritic porous bypersthene-amlesite with porous ground- mass, built up of colorless glass with n, felt-work of augite-niicroHte and a little skeletal plagioclase. 2) There is a shallow bank, 1 km. w-estward from Yébisu-jinia (Nos. 6-8). The writer covild not find any description that corresponds in position to this shallow. This TJoro-shhnd ('mud island') or Kani-shima (Fig. 11) is a submarine pumice bank of yello\\ish- brown color, 436 m. (4 clio) in circumference, which appears above water only during lo^\- tide. This may be the phantom island, kno\\n only by traditition, which is said to have api^eared in the Tembio- Hôji era (764 a.D.), and lately referred to as Ko-jima (pp. 2'J, 41). An old book says, according ti) Yamaguclii, that an island appeared on October IS, 1781 at the west of the new islands only for a few days. Again in June or July, 178C, another one rose at the west of the former and soon disa])peared. The shallows to the f.outh of Doro-shima îire saiil to be the positions oE these short-lived islands. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 53 (ceramicite) thrown out from the Nabé-yama vonts during tho Ih'st phase of tho recent volcanic activity. orioix of Grlijin of the new islands.— The foregonig pyrogemc New Islands islaiids iiiado thcii" appoaraiicc as the outcome of at- tendant submarine effusion of lava during and after the eruption in 1779. The following views as to their origin may be men- tioned : a) They are sometimes spoken of simply as vents on the rupture line that traversed the body of Sakura-jima, as the Nabé-yama vents in the last eruption, b) It may also be suggest- ed, although no one seems as yet to have given utterance to this view, that they are the loci of ventholes on tho front of the gas- pent and cooling submarine lava that plunged into the sea down the mountain slope in a fiery stream from the twin vent, the east and west Idgona or ' fire -holes ' near the northern top (Fig. 11). Neither view is applicable here as a working hypothesis, if the bathymétrie condition of the surrounding sea be critically examined. The new-born islands occur in diffuse manner and rise preci- pitously from the bottom GO fathoms deep. Local depressions of the same depth intervene between the shore and the islands, so that there is no rooni for doubt that there exists no real connec- tion of lava- flows between the insular masses and Sakura-jima (Text-fig. 11). It seems to the writer that they represent the positions of ventholes of what Wolff '^ styles the central eruption upon a local magma reservoir of satellitic injection fed through a narrow chan- nel from the main macula under Sakura-jima, just as, according to Daly,^^ Kilauea is fed by Manna Loa in the Hawaiian islands. The 1) F. V. "SVolff, ' Der Vulkanismus. ' Bd. I. 2) 'The îs.iture of Volciinic. Action.' I'roc. Am. Antd. Arts and Sciences, Vol. 47, 19D 54 ART. 3. — B. KOTü : bottom, 60 fathoms deep, was under pressure and comparatively cool. The feeding channel from the main was closed by consolida- tion, and the accumulation of liberated gases in the cooling secondary reservoir under high tension corroded and ate away the roof, finally- opening the way for the access of water to the small hot macula. Soon subaqueous outbursts followed, ejecting mud (fine splin- ters of pumice) and lapilli in the shape of fountains, and dis- turbing the body of water which caused tsunami or sea-waves to the surrounding shores. The shallow bank, outlined with a 10 -fathom line, rose from the bottom by accumulation of the sub- marine éjecta (Text- fig. 11), and likewise the lapilH-ash bed of Shin-jima (No. 5), 44 m. high, was deposited and raised at this time. As in normal development of submarine islands, a massive dome or tholoide rose through the aqueo-pyroclastic deposit in the final phase, which is typically represented in the form of Iwô-jima (Nos. 3-4). One of the writer's motives in visiting these islands was to see the products of subaqueous consolidation of lava, the modus operandi of which is exactly known from authentic records. In contrast to subaërial lavas, compact submetallic- clinking lavas were not found in the surface rocks in the whole region over which the writer rowed. The rocks are all more or less pumice- ous and sometimes slightly slaggy, they have the appearance of a dusly coke from lignite and are also light. Under the microscope the intratelluric plagioclase-phenocrysts are compound crystals witli brilliant birefringence, and the crystals are full of glass due to temporary regressive and corroding action of the magma during crystal-growth. Hyporsthene and augite are as in terrigenous lavas. In the glassy groundmass one finds only slender pyroxene with stiff fibrous terminations in brown glass. The latter contains THE GREAT ERUPTION OF SA.KURA-JEMA IN 1914. 55 abundant globiilitic bodies, and is often locally decolorized, ap- pearing macroscopically as light flecks. The pyroxene may be mostly enstatite-augite as in artificial slags. Magnetite and feldspar- microlites are wanting in the I'ocks from Inoko-jima, while in those of other islands magnetite-dust and skeletal plagioclase- microlites occur in moderate quantity. The rock is cloddy, due to contraction-fissures variously traversing the whole body, formed by a sudden quenching of the hot mass in water, the clefts so pro- duced are afterwards fined with a limonitic substance, and filled with black dust which marks the direction of flaws. Some of these characteristics differentiate the subaqueous from the lavas of subaërial consolidation. pha^e^o "^^^^ great paroxysmal eruption of the An-ei erfi AcTrïSY (1779-80) echoed for many years. 1781 On April 11th, 1781 (10th year of An-ei, or 1st year of Temmei era), an eruption took place (where ?), and the people in the island were greatly frightened. They made a vow to the God of the mountain, and refrained from fishing on the 18th (lunar calender) of every month. This day they still keep. While the writer was in Kagoshima last year, he heard people talking about this sacred 18tli, whicli had lately not been strictly observed by the inhabitants, causing tl^e wrath of the God Gongen whose shrine is located on a fittle elevation (400 m.) on the western slope, and the principal centre of the late eruption on the western flank was near the said shrine now entirely blown off with the hill. 1782 On January 18th, 1782 (2nd year of Temmei era), the mountain Was active ? 1783 On September 3rd, 1783, it sent out flames, and ash fell as far as Kyoto. 56 ART. 3. — B. KOTÔ : 1785 On November 20th, 1785, it again erupted. 1790 On July 29th, 1790 (2nd year of Kwan-sei era), the top was much shaken and flames came out. 1791 On September 11th, 1791, it again bm-nt. 1792 On October 11th, 1792, it again burnt. 1794 It erupted. 1797 It sent out flames. 1799 On Marcli 27th, 1799, it sent out flames. Eruption- Duriiig the year 1835-'36, a hot-sprimr in Arimura PEKIOD Vl. l=> J i O 1835-1836 gushed (cf. 1779, p. 47). I860 In February or March, 1800 (Ansei or Mannen era), the Yamano-yu hot- spring (?) in Takaguma, and another, a luke- warm sulphur spring in Tobi-oka on the north of Tarumizu, both A in Prov. Osumi, gushed out. The south crater of Sakura-jima was rather active. 1876 In 1876 (9th of Meiji era), a hot- spring in Shin- midô (?) in Prov. Osumi gushed out (Milne). 1878-1879 At the end of the winter 1878, and in March, 1879, the south crater sent out flames (Friedlaender). §111. The Eruption of Sakura-jima in 1914.'^ (Eraptioii-periocl A-TI.) 1913 A) Premonitory Symptoms. —As early as the spring 1) The following fire the chief "works concerning the recent ertiptions : Nishio-Friedlaencler, ■ Der verheerende Ausbrach des Vulkans Salaira-jima im Süden der japanischen Insel Kiuschiu.' Feternvmns Geogr. MUteil, 1911, S. 132. It is an extract from the ' Explanatory Text to Sheet Kagoshima.' Oseki, ' Der Viilkanaiisbrach anf Japan.' Zeitschr. Gesell, f. Erdkunie, 3'« Berlin, 1914, S. 151. Yamasald, ' Der Ausbruch des Vulkans Sakurashima im Januar 1914.' Zeitschr. Gesell, f. Erdkunde m Berlin, 1914. F. A. Perret, 'Preliminary Eei)ort on the Grand Eruption of Volcano Sakurashima.' (Ibilian) Zdlsdir. f. Vulk., Bd. I. S. 113. THE GIÎEAT ERUPTION OF SAKURA-JIMA IN 1914. 57 of 1913, Kagosbima and lier environs showed many precursory and abnormal signs of subterranean convulsion. During the spring HoT-spRixGs of 1913, the hot-springs of Kurokami on the eastern coast of Sakura-jima ceased to flow, so that the inn-keepers were obhged to give up their business. A little later, at Arimura and Yuno on the south, tlie alkahne hot- springs, all situated at the water's edge, became too hot for bathing during low, l)ut cool during high tides through dilution with sea water. These incidents indicated that gases were already ascending from crevices, or lieated waters seeping down from above. {Sec p. 47). From the 19th to tlie 30tli May, especially the 22nd and tlie 23rd, the atrio-hke hollow of the Yoshimatsu area, "^'eÏeÎh^*^^ ^'"^ ^''^'^- '^o ^^^^ north of Kagoshima, with the epicentre '^"'^^ ' near the above-named railway station at the northwestern foot of the Kirishima volcanoes quaked repeatedly, accompanied with detonations {sec the Summary), causing anxiety to the in habitants. On June 29 and 30, while the quakings of Yoshi- matsu Avere attaining their climax, sympathetic earthquakes tem- porarily occurred near Kagoshima and the westerly lying pumice T. A. 'Tagger Jr., ' Letter from Sakurfi-jimu.' Weekly Ball. Uaicitiian Volcano Observatory, No. 13, 1014. Omori, ' Tlie Sakani-jim.i Eruptions and Earthquakes.' I. and H. Bull. Imp. Earthq. Invest. Com., Vol. Vni. TS[os. 1-2, 1914-1910, Tokyo. Satf), ' Eruption of the Volcano of Sakurajima.' Ball. Imp. Geol. Sarv. Japan, Vol. XXIV. No. 1, 191 1, with an English resume. Azumfi, 'The Great Eruption at Sakura-jima of 1914.' (:h iE '=. ^ tM ai Tk. \\U '/< hB) Kagoshima, 1914, pp. 1-2.52. (Japanese) ' The True Account of the Great Explosion of Salaira-jima in 1914.' {-}^ jp. ;-i. ijî. @ J^ ;^ 'Ig ^ ^ nû) Published by IVie Kylshû Daily News, Kumamoto, 1914, pp. 80. (Japanese) ' The Great Explosion and Earthquakes of Sakura-jima.' (f§| ^ ^ 'J:|4 S ffS) PubHshetl by the Répertoriai Staff of The Kaqoshima Nevis, Kagoshima, 1914, pp. 340. (Japanese) ' Sketclies of the Great Eruption of Sakura-jima in 1914,' by the Meteorol. Observ. Ivago- shima, 1915. :;^ E H ip- ti â Hi :^ nR 'A ^ S Hi Official Eeports of the Kagoshima ^Meteorological Observatory on the recent eruptions are found in Jour. Meteorol. Soc. Japan for 1914. 53 ART. 3. — B. KOTO: region of Iziû-in, 15ä;;?z. west from the city, where clocks stopped and a few houses were badly damaged. On November 8th and December 9th, 1913, and on the 1914 8th of January, 1914, Kirishima, lying 6o km. to the north from Sakura-jima, made successive explosions, ejecting ash over the surrounding districts. A week before the great catastrophe of the 12th, a pond 3.6 ??2. square, just above Ari-mm^a in Sakura- jima suddenly became dry, and the fishes in it all died, and some wells in the island dried up, or the water-table became con- siderably lower. These unusual phenomena were reported and the authorities replied that these were mere effects of the activity of Kirishima already referred to. These movements embracing the region within the radius of (31 km.^^ from Sakura-jima already pointed out the local origin of a subterranean disturbance of great volcanic energy. On January 8, thick snow fell in Kago- shima Bay, which has been rarely experienced in warm southern Kyiishii. ^^^ jQ^^ We were now approaching the dreadful crisis, the ■^^^* energy concentrating in Sakura-jima in its upward ex- ertion for the egress of lava. From the morning of January 10th, precursory shocks were experienced in Kagoshima, throwing water out of basins every (?) 10 minutes. From 7 p.m. rapid tremors shook the island, accompanied with subterranean groaning. The magma was then apparently ascending. Some people observed already during this day shaken rocks falling from the western slope of the north cone of Sakura-jima. The night was stormy with flashes of lightning followed by rain, then the weather became clear and warm. 1) CI. K. Gilbert, ' Interpretation of Anomalies of Gravity.' Prof. Paper 85 C, U. S. Geol. Sm-r., 1913. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 59 •rAx. iiTH, Qii the nth, 3.41 a.m.,^' a weak local earthquake Morning * (scale 1)'^ was felt in Kagoshima, which inaugurated the sub- sequent convulsions."'^ Henceforward, feeble shocks occurred four times an hour on an average till 9'' 51' 45" a.m. when a strong one (scale 5) shook the city. The direction of the earthquake vibration was from S.S.E. to N.N.W., or S.E. to N.W., as recorded on a Gray-Mihie-Ewing seismograph in the Meteorological Obser- vatory in the city, situated on a hill-top (120 m.) of a pumice bed 90 m. thick. If the seismometer had been posted on solid rock in a region of flat topography the writer conjectures the direction of vibration would have been somewhat different from that actually recorded as above stated. Just at this time a boy at Nojiri in 8akura-jima was hurt on the head by flying stones (explosion ?). About four hours before the earthquake of 5.58 a.m. (scale 2), a slump*^ of considerable magnitude happened at the doubly pointed ' scissor-rock ' or Iiasami (Geologic Map) on the western crater- margin of Mt. Sakura-jima, which slid down toward the point The where vents opened later. The slip caused mighty Avalanche OF THE avalanches of debris and dust clouds which might Have SCLSSOE- EocK been easily mistaken for an explosion. The fresh, rusty- 1) The following is chiefly an extract from ' Report Meteor. Station Kagoshima.' Jour. Soc. Meteor. Japan, No. 2, Tokyo, 1914. As to the numerical tlata refer to the original Report (see also Summary). My creditable informant is a fisherman who passed the restless night of the 11th -with the vilLagers at Nojiri in Salaira-jima. He fled for his life at the explosion of the 12th at 10.5 a.m. on his lioat to Tani-yama on the Kagoshima side. Then, rowing northwards along the city, met on tlie way homeward the great earthqiiake at 6.29 p.m. on a turbulent sea, and finally reached Fulmyama in safety the next morning. It is this resolute man who offered his services to guide the -writer on his trip on the 18th January along the coast of Ôsumi as far as Fumoto near the lava end. He knows the details of the catastrophe from the very beginning to the end. 2) The scale is the one adO]jted liy oiar seismologists and current in use among us. 3) This earthquake startled the whole population of the city from sound sleei) ; from then on they were kept awake throughout the early morning by the ever-increasing tremblings. 4) The dust-clouds formed from avalanches are not easily distinguishable from the ash- clouds of real explosions unless actually observed from proximity. 60 AKT. 3. — F.. KOTÙ : brown scar loft by the slip can be distinctly seen from a distance, and tlie disappearance of pointed rocks deformed and lowered the topography of the western slope. The 11th, ^^^ earthquake of the same intensity (scale 2) was Afternoon • i t i -i -, i -i experienced between noon and 1 r.M., and some people saw a light-bluish cloud rising on the western slope (the first sign of explosion on the west?). About 2 p.m. a man in Tarumizu saw light white clouds rising just above Ari-mura at about what cor- responded later to the Sen-yemon venthole (No. 1 vent) on the southeastern slope of Mt. Sakura-jima (the sign of explosion on the cast?). It may have been either condensation of steam below, or of vapor in the air by ionization. This fact indicated that the lava already ascended through the channel up to within 300 m. from the surface, as it has been experimentally proved that the liberation of gases begins with the pressure corresponding to that depth. From noon, in fine weather, the frequency of the shocks was 10 in an hour till 8 p.m., whence it doubled, keeping this up till about 10 a.m. of the next day, the tragic 12th. The recorded shocks from the beginning down to this moment (12th, 10 a.m.) were 418. The 11th, Shoclvs of the samc degree (scale 2) occurred once between 5 and 6 p.:\i., ahd again between 7 and 8 p.m. At 7.15 p.m., cannonade-like sounds and tremblings occurred for The West the fii'st time. A policemau at Yokoyama in western Sakura-jima reported at night to the authorities that high up in The fi.yino the back mountain he saw smoke rising and red-Jiot OF KeD-HOT Stones stoucs ßjjinfj (ihout. Duriug tlic wholc day the island was incessantly shaken, and roof tiles slipped from the eaves, showing the peculiar nature of the quaking. Apprehending the THE GREAT ERUPTION OF SAKURA-.TIMA IN 1914. 61 coming dreadful event, a large proportion^^ of the aged, infant and female population of the island fled to the nearest mainland from early day till lat(^ in the moon-lit night of the 11th. The whole population around Kagoshima Bay spent a restless night, especially those who did not take to their heels from Sakura-jima passed the ever-trembling night in the cold open air. Clouds of gases were no doubt already issuing somewhere from the western slo}>e during the night. The East Of the cast WO kucw vcry little, as the inhabitants had already deserted the island during the day, only a very few invalids and policemen'"^ stayed behind. B) Surface Manifestations of Activity at Sakura-jioia.— The rjijj^, FiKS'j' Phase. Eventuallv the fateful dav''^ dawned ; ^ rira^ ^^^^ weather was fine and calm ; the low pressure had . oRNi>G si^iftod its centre eastwards already on the 10th. The quakings abated a little between 1 and 3 a.:m., suggesting that during these remarkably quiet hours fissures or spaces . opened 1) 2i,577 out of 27,lir, of the iiiliabitants in the villuges with which the island is dotted all round the coast. 2) The Avriter knows of one case, where sufferers at Jvurokaini were literally entombed under the pumice-lapilli burden (2 m. thick) and in their crushed houses for two days without food; at last they took courage to embark on a boat which was entrapi^ed in the thick pumice sea {12 cm. thick). They were saved by some volunteering yoiiths, who, noticing handkerchiefs waving from a distance, hastened to the spot and brought these poor i^eople to their own homes at Fukuyama on the 14th. 3) A Japanese author said 137 years ago, when he wrote about the An-ei event at Saki;ra- jima, that eniptions h.ibitually took place at New or Full Jloon. That of An-ei was at New ]\Ioon— the first day of the lOth month, 1779, while this year it was near the Full Moon, i.e., the 17tli day of the 12th month of the lunar calender. Jensen says that A'olcanic eruptions occur more frequently at Full Moon and New Moon than in the Last quarter, and at the end of sunspot maxima. ' The Geology of Samoa and the Eruption in Sawaii.' Proc. Linneau Soc. XeiD South Wales, 1906, Vol. XXXI. part 4, p. 665. By the way, it is to bo noted that P.J. Eicard, of Santa Clara College, Cal, lately observed a sunspot of great dimensions, twice the area of the earth, i.e., 1/278.5 of the sun. Die Umschau, •Taniiary Number, 1914. 62 AKT. 3.---B. KOTO : for a moment for the influx of magma into them. The earthquakes revived afterwards till noon ; the climax, however, had already- been reached during the night before. The insular population felt the strongest shock at about 5.30 a.m. Then surface activity was drawing closer. At about 8 a.m., white filaments of misty clouds were seen to rise after an earth- quake near the Gongen shrine^^ at an elevation of 300 m. on the west slope facing Kagoshima, while at about the same time the Hot and saliuc hot-spriug, and wells at Ari-mtira on the sou- Springs tlicm foot spoutcd iu fouutaius to a height of 1 m. It will be remembered that this village was afterwards buried under ' live ' lava. The cold mineral spring at the water's edge in Saidô on the northern shore spouted at about 8.30 a.m. As the spot is known by the name of Yuno-saki or the ' hot-spring point,' it is likely that formerly the spring possessed a high temperature. The outgushes of this and other springs all round the island may not necessarily be attrilnitable to subterranean pyrogenic origin ; on the contrary, the writer is rather inclined to assign the cause to disturbances of drainage by the constant shaking and trembling high up in the body of the volcano. At 9.10 A.M., a thread of white steam clouds (Text. -fig. 14 a) rose upright from the southern top-crater (Minami-daké) where two weak solfataric vents had l)een constantly emitting vapors from the inner cliff ever since the activity of 1879.'^ It was a sure sign of surface activity, as the lava canal is directly connected with the crater. In the An-ei eruption of 1779, the same pioneer erup- tion happened in one of the three top-craters. {See Fig. 10 a, p. 44.) About five minutes later, an earthquake (scale 3) happened associated with cannonading sounds and tremblings, and then the 1) Near No. 1 vent in Geologic M.i]). '2) See p. 56. THE GlîEAÏ ERUPTION OF SAKURA-JIMA IN 1914. 63 Fig. 14. — The tirst i)bHse of éruption, viewed from Fiüaiyam.i.i) western (Yuno-hira) vent in calm weather threw up black curdy smoke (ash cloud) high up into the air, which expanded in tlie upper horizon in cauliflower shape (Text-fig. 14 a (2), h on th(^ right), meanwhile red glares and sparks were emitted from the vent, signalizing tlie advent of fluent lava. The eastern (Nabé-yama) vent a little later sent forth clusters of black clouds (Text-fig. 14 6 (3)) accompanied with terrible de- tonations. Tims, the general sur- face outbreak was in- auijurated on both sides of Mt. Sakura-jimaP West Side r^^^^ ACTIVITIES ON THE West Side. Once the out- ])reak had started, it seemed to know ]io bounds. At 10.40 a.m., explosions and detonations were outbkeaks Inau- gurated 1) Fig. 14 is copied from a hand-sketcli by Doctor Ninomiyii, of Fulinyuma. 2) "When at last (10. 10 a.m.) Mt. Sakiira-jima was universally recognized as the real seat of the volcanic disturbance, those people who still remained on the island took flight on small craft still left on the coast, or were waiting half mad on shore for succor. All available steam- launches were speedily sent out from the city to all i)arts y tlie Eailway Station-master of Ivigoshima. The height calculated from Kushino, far away -west from the city was 7,272 m. ; that estimated from the man-of-war "Tone" anchored near Taniyama being 8,181?;). T. Ogura, my assistant, got the value 3,560 77). from the writer's photo of January 15th, i.e., four days after the first outbreak. This was, the writer presumes, the average height of clouds for a few days while displaying the Strombohan type of activity. In the Report of Meteor. Oh.ierv. K'Kjoshirna, it is stated to be 3,000 m. The smoke culumii of the Krakatoa eruption was estimated as 27 km. (17 miles) high; in the Mauna eruption of 1911 it was 8000 ft. (2,401 w.)— Wood, Jiull. 6'ei.9. Sac. Am., Vol. V. In the An-ei eruption of Sakura-jima, the height calculated was 12,029?».. (see p. 45, footnote). •2) Zeltfidrr. f. Vulkanologie, Bd. I. 1914, S. 25. 66 ART. 3. — B. KOTO Holes in Ground Numerous holes (PI. Y. Fig. 2), evidently caused by trajection of large disrupted blocks, were to be seen on the western shore, varying in size from 0.5 to 6 m. in dia- meter and 0.3 to 2 m. in depth. Shattered blocks with cordierite- bearing white inclusions are still to be seen in the bottom of pits made by them, or scattered about in the neighborhood by their rebound. Electric discharges seem to have had nothing to do with the making of the hollows. On the other hand, it seems not improbable that a thick pumice -ash deposit on the ivet valley bottom might have been puffed np by steam, generated hij the advancinr/ and overriding fluent lava, and in this way some conical pits may be formed in the ground. The 12th, From uoou, tho earthquakes remarkably abated owing .\fteknoon ■"■ ./ o to the releasing of pent-up gases, which opened a way for the lava to the surface ; but ' air-quakes '^^ or air -concussions and roarings con- tinued, increasing both in frequency and force, con- stantly rattling the windows even of 1) The air-cxuake or (/«(«.si-earthquake is accompanied ■with the ' cletotiatioti ' — the air-wave, which in contrast to the pure sound-wave is said to be i)roduced by slow oscillation of air. It does not affect the human ear as a sound in itself, but it causes audible sounds when it encounters some obstixcle which can be easily disturljed, such as Japanese sliding doors. See S. Fujiwara, ' Abnormal Propagation of Süv;nd-waves in the Atmosphere.' Bull. Ceritr. 76. The rate of flow Ijecame ikiw greater. 80 ART. 3. B. KOTÛ : was again in the city, which was then restored to its former state, and was fall of life in every way. East Side Nothing was known of tlie details about the Nabé- yama vents. The prime of the Sakura-jima eruption after having been active for nine days, was now near its end, especially with regard to the western vents. Jan. 20ih ^he Yuuohira lava was slowly pushing on with white fumes forwards into the sea, having already drowned Karasu- jima island (PI. VIII. Fig. 1), while the vents were tranquil, roaring periodically, though the trembling ceased. The cliff of Atago- yama, 167 in. high, on the back of Akamizu was still free from lava ; but its southern prolongation was already overflooded with melted rock. See Geologic Map. East Side ^\^q writer took a Small steamer to Tarumi in Osumi to get a sight of the Nabé-yama vents from the south to complete MiGEATixG the round of Sakura-jima. The steamer steered heavily SHEET through the migrating pumice sheet (Text-fig. 21), which, as already referred to, the writer met with on the 17th on the northeast of the bay (p. 81). The whole stretch of coast from the west end of Arimura to Séto for a distance of 3 hn. was entirely buried under a lava sheet and enveloped in impenetrably dense fumes and clouds. This side was very active in contrast to the west, constantly can- nonading and vibrating both the ground and the air. PL X. I'igs. 1-2. It was somewhat astonishing to find that the south -«^esif. coast of the island between Yuno-hama and Nojiri remained intact. No damage to houses ; no pumice on the ground. The ash covering was, however, universal. The region in question is the land neicly- huilt by lava flows in historic periods. The present enijJtions and THE GEEAT ]:RITTI0N OF SAKURA-JIMA IN 1914. 87 hira flous carefully (iroidcd this Si'dion of Sohira-jima. Even the temperature of the hot-spring in Fiirusato was not greatly affected, l)eing 45^ C. which was only 2^^ C. liigher than in former clays {see p. 47). Jan. 21st fpjj£ Thikd Phase. The sccond phase of the eruption (luring the interval from the 14th to the 20tli was now closed. The city side was quiet ; lava still moving forwards into the sea. The writer took the train on tlie 21st for Tokyo, so the diary for the following days is not from his personal observations. It is said that from 7.5 p.m., a strong explosion and roaring occurred, continuing throughout the whole night. The sea glared from the lava flow. Ash fell on the city. About a thousand people again fled to the west from the city. Nothing was known of the east side. But it nmst have been more active than on the west side, throwing pumice and ash A to the opposite coast of Osumi. -Tan. 22Nr) r^\yQ yr^iYv of asli OU tho city coutiuued from last night (118 grams per ^(\. m, 2 mm. thick). Day^^ and night there was constant roaring. Nothing particular was known from either side. The lateral lava margin was said to have been pushing forwards one foot'^ an hour near Hakamagoshi on the west. Jan. 23ed From midnight it began to roar, giving a sound like a landslide. This was probably caused by the widening of pits on the east side. The citizens passed a restless night ; the afternoon was misty, due to floating ash (21.5 grams per sq. 772.) ; a strong earthquake occurred at 9 A,:\r., the day was rather quiet due to the outpouring of lava, but at night it again roai-ed. 1) About the sjk titer »■rupthii'.s observed on this day the reader is referred to tlie SumriKiry. 2) Vldfi ante, p. 76, footnote 2. 88 ART. 3. — B. KOTÔ : The Nabé-yama lava was fast moving under water, but the heads of the lava-front rose up and formed a fuming series of equatorial islets parallel to the stream-line of the 8éto channel. The venthole No. 4 (Text-fig. 19) seemed to have fpvicn in sue, this, the writer thinks, was due probably to the ash ejections of geeatLata fj^f, i^fgi three days. The Java fiou' then received ahund- FLOWS IN ^ yama'^iTnts ^^^^ supplies from the widened pit. It was on this day that the lava-free, cliffy Akashi Gongen hill (103 7??.) between Waki and 8éto became completely overflooded (PI. IX. Figs. 2-3, PL X. Fig. 1, Geologic Map), and the lava tongue around Nabé- yama pushed considerably forwards toward Kurokami. By morn- ing, the channel, which was formerly 400 m. wide and 8-0 fathoms deep, was open only about 24 m. The lava flow on this side was so rapid that on the next day the channel became almost land-locked. A Ash fell here and stifling gases swept the opposite coast of Osumi. It was the first time that the writer heard of suffocating (jases in the present eruption, indicative of the declining phase of activity. Jan. 24th r^\^Q activity was gi^adually diminishing. During the morning it roared periodically and ash was still precipitating ; only white fumes were seen on the city side ; the lava moved little by little into the sea (29^ C.) with saline fumes. The east side was also rather quiet and exhausted, though enveloped in white clouds. The channel of Séto remained in the same condition, though becoming a little narrower. PI. X. Fig. 1. Jan. 25th rpj^^^ activity bccame feeble, the Yunohira vent being apparently dead, and the lower ones were only making little secondary eruptions. The east side was, ho'v\fever, still roaring and exhaling, and the whole area was entirely enveloped in vapor. Little ash fell, and odious gases swept down. THE GEEAT ERUPTION OF SAKURA-JIMA IN 1914. 89 ■Tax. 20th rjij^^ activity was fast declining. It roared at 7 p.m., 11.42 r.ii., 11.51 P.M., and at 0.15 a.m., 0.42 a.m., 1.23 a.m., 2.15 a.m., and 5.50 a.m. of the next day. The lava was moving a little under the sea. The east was active, throwing up smoke to a height of 2,000 m. The entire region was wrapped in white fumes. The Séto channel was still open (18 m..) though becoming narrower by 6 m. One could not row through the boiling strait (60^' C). .Tax. 27th ijij^q activity of the Yunoliira ventholes passed aicay with this day, though that of the Nahè-yama craterlets still lingered on fo>' 6 days. Feebly active in the Yunohira area, with occasional minor secondary eruptions and ' trombe.' The entire field had a melancholy aspect (PI. IX. Fig. 1) like that of the hot rubbish of Apparent ^ \yjcr towu after a great fire. The lava-flow in the sea Movement ^ ^ laTa^IIo^v ^^'^^ ^^^'^ at a standstill before the city of Kagoshima, siAXDSTM. the new lava -field, 1.4 sq. km., extended into the sea 950 ?7i. from shore with a breadth of 1^ hw. (Text-fig. 19). The Séto channel on the east was agitated. It was now re- duced to 9 m. in breadth, and 3.6 m. deep with a brown pumiceous bottom, observed at 3 p.m. at low tide. The sea w^as hot, but only for the upper 90 cm., while at lower depth the water was cold. Jan. 28th rj:^^ FouRTH Phase. 27ie third p/irtst' of eruption ex- tending from the 21st to the 27th icas now closed, and the activity entered into the fourth phase. Tremors occurred only at 5.49 p.m. ; roarings w^ero heard IG times in the city from 4 p.m. tiU 10 a.m. of the 29th ; glares of lava at night became weak and were seen in 5 spots. The lava of the Yunohira vents apparently moved no more. On the east side, the channel was still not yet closed by lava wdiich was moving fast forwards, being 9 m. in height on the 90 • AKT. 3.— B. KOTÙ : lava-front, and 7 m. deep at the passage at low tide. Three men crossed through it. Constant shocks were felt during the night at Fumoto near the strait, keeping the people awake. Jan. 29th r^\^Q activit}^ on the city side practically subsided with occasional roarings. Earthquakes were, however, recorded in the city at 6.43 a.m., 11.42 a.m. (scale 1), and 11.41 p.m. (scale 2). The east side was still active, constantly pouring out magma. When a party from the Fukuoka University approached tlie channel, they found it practically closed by the collapse of the IS m. high lava- front, caused by a secondary explosion at 2 p.m. Jan. 30th- Nothing particular happened on the 30th. On the 31ST 31st, 0.30 p.m., an explosion occurred on the west, by which the city was air-shaken. Strong smoke rose on the east side. Feb. 1st -^q^ fQ^P days the Yunohira ventholes remained quiet. On the other hand, at 5.43 a.m., an earthquake was recorded in the city (scale 1). At 2.17 p.m., 4.20 p.m., and 7..") G p.m., roarings were heard with sharp tremblings and ejections of strong smoke. The Nabé-yama craterlets were still active and a new one was opened somewhere. Apparently a new islet appeared near Ari-mura, separated from it by a narrow water. This afterwards proved to be a lava head risen from the shore bottom. At the time of the earthquake at 4 p.:\i., a strong explosion occurred, the lava advanced, completely choking up, according to Mr. Schwartz, sakura- the strait of Séto. This the island of Sakura-Jima JIMA A Peninsula became permanently a peninsula attached to the province of Osiimi (PL X. Figs. 2 and 3). Feb. 2nd q^ ^^q wcst, it strougly cannonaded and shook the city windows at 2.5 a.m., roared at 8.30 p.m., and 8.43 p.m. ; lava glared 3 or 4 times during the night. The Nabé-yama side was THE GDKAT EliUPTION OF SAKUItA-JIMA IK 1914. 91 increasing in activity ratlicr than declining ; the smoke, however, became less in qnantity, it roared strongly at 0.05 v.M., 7.15 p.m., and 10.50 r.:\r. Feb. The Fifth Phase. The tburtli phase of activity, inckiding the period from January 28th to February 2nd, was now closed, and the status entered into the fifth i)hase. This, the declining phase, still continued, manifesting the after- eflects of the greatest volcanic activity we have seen since the An- ei eruption, 130 years ago, in volcanic Japan. The activity on the east side came to an end with the day immediately preceding, and Maix vul- at the same time the main mdcanlsm of Sakiira-Jima, the CANISII Ol" sakvba- citv" side of it having been already at rest since the 27th. JIÎCA TEBMI- " cj ^ NATED rpj^g, minor manifestations w^ere mere after-effects. Tranquil on the v:est during the day, only making minor ex- plosions 2 or 3 times in an hour. There was a lava glare only once during the night and an explosion at 0.10 a.m. The east was still active, ejecting ash late at night and early next morning, which fell in the Kimotsuké and So districts in Osumi ; a strong detonation occurred at 9.57 p.m. The lava-fronts w^ere pushing seawards little hy little in both arms. Feb. From early morning posthumous smoke rather in- creased on the west, 4 or 5 times every hour. Of the east nothing was known except occasional air -concussion s which shook the city. On the 0th, 9 a.:\l to 11 a.m., air -concussions from the east side rattled windows in the city at 7.55 p.m., a bright light was noticed with outpourings of lava. An earthquake (scale 1) occurred Feb. ^^t 11.48 P.M. 7tii— 8th At 4.40 a.:m., it roared rather stronglv and trem- 92 ' ART. 3. — B. KOTÙ : bled on the city -side while one or two small explosions occurred every hour. Ash fell in dusky weather on tlie city from G.30 a.m., due to an explosion oa the east (where ?). Clox-d- On the morning of the 8th, a cloud-]3Urst swept BtTßST AND ^ Mud-flow f^^Q coast of Osumi, causiug a pumice and ash flood in Ushiné village submerging 400 houses. It is to be emphasized that the weather had been fine since January 10th, liitherto crea- ting no mud ßows (see Summary) in the asli-covered area during this rather long interval, excepting little local rainfalls. It is also to be observed that these precipitations, slight or strong, always happened after an ash day, but not simultaneously with the ascent of steam columns in the main vents. Feb. The morning was fresh. The western side was calm and only smoked once at 3.25 p.m. The eastern was strongly smoking at 1 p.m. On the lOtli, nothing was record- ed. On the 1 1th, no smoke w^as witnessed on the west ; the lava front was still emitting saline fumes in the sea. The east was yet strongly smoking and roared from 6.42 p.m., wliile the west Avas exhaling only faint smoke on the 12th. Feb. The east roared at 7.20 a.m., 4 p.m., and 6.41 p.m., the 13th-14th , , last one was rather strong and rattled the city-wmdows. The icest detonated strongly with smoke at 6.41 a.m., became bright at 11p.m., and roared at 11.44 p.m. Earthquakes w^ere fre- quently felt at 3.17 a.m., 8.52 a.m. (scale 1), 3.21 p.m., 3.34 p.m. (scales 2-3), and 6.30 p.m. (scale 1). On the whole the ground was by no means tranquil. iwô-jiMA It was on this day, the 13th of February, that Eaeth- tiTJAKEs the active volcanic island of Iwô-jima (pp. 5, 21), the " Sulphur Island " (Fig. 1 i. Fig. 2), lying 100 hn. southwards off the southern shore of the province of Satsuma, was strongly ÏH]-; GEE AT ERUPTION OF SAKURA-JIMA IN 1914. 93 shaken iVom 2 a.:\i. to 3 r.'siP, causing anxiety among the inhabi- tants. It is an island 20 km. in circumference with a population of about 000, of which 300 are laborers on the sulphur mine. For twenty years the volcano had remained comparatively (|uiet. There is a single main cone, 770 m. high, besides 2 parasitic ones. Friedlaender' says, that there are fumaroles and solfataras in the crater, 7-800 m. in diameter and 50 m. deep, as well as in a deep glen on the southern slope. It may not be unreasonable to con- clude that the earthquakes, above referred to, were the manifesta- tions of vulcanism^^ in the zone to which this island and Sakura- jima as well as Kirishima are the appendants (•'^6^6^ Text-fig. 2, p. 11). (3n the 14th, only the eastern vents, as usual, strongly smok- ed and roared ; red-hot lava at night glared at the vents. Feb. The east boomed several times and ash was blown toward the city, turning it snow-white, and Sakura- jima was veiled in ash-mist. Again it rained fi'om 3 r.:M., wliich washed off the dust of the city. At Tarumi in Osumi a freshet Second of ash-mud broke dams and wrought havoc in the delitge village. This was the second ash inundation (p. 92). On the 16th, the west was still faintly fuming. The east was, however, roaring, and an air- concussion at 10.58 p.m., rattled the city windows. Slight ash again fell from 4.40 r.:\i. Smokes were low on the east, while the west Feb. i7th-i8th temporarily woke and roared, and sent out flames at 128 a.m. and white clouds at 7.45 a.m. On the 18th, slight de- 1) Distant earthquakes were recorded in Kagoshima from tlie 13th, 3.31 a.:^[. to the l.jth, 5.13 A.M., of which the first was the strongest. 2) See ante, page 21, footnote (a). 3) See page 98. This illustrates a case of ' sympathy of alternation ' among the vol- canoes belonging to the same zom (not the regions). See T. A. Jagger, Jr., ' Activity of Mauna Loa, 19U-la: Jmrr. .Tom: Sn., Vol. XL. 191.5, p. «39. Feb. 1!)th-22nd 94 ART. 3. B. KOTÙ : tonations at 0 a.ivi. and 9.30 r.M. on the east. It rained again from 9 A.M. Misty clouds covered the island. Air-vibrations from the east rattled the city. The west emitted smoke at 7.50 a.m., and also 11.30 a.m. On the 20th, tlie vrest roared at 7.30 a.m. Mists veiled the smokes of the c^ast. On the 21st, 7.15 A.M., a great detonation from Sakura-jima, which was enveloped in rainy clouds, alarmed the city population. ()n the 22nd, the west roared at 10.5 p.m. and 10.43 vm. For four days the island was hidden by mist. „ On the 23rd the east roared at 2 a.m. and 6.49 a.m., Feb. ' 23Br)-24TH ^^^^^ ^^^^ ^YiQ west at 4.56 a.m. An earthquake was felt at 9.18 a.m. Slight ash fell during the whole day of the 24th; a rather strong booming was heard at 0.44 a.m., and 5 times from 10.54 A.M. j,^^ At 2.41 A.:\r. loud roaring ; the city windows rattled 25TH-26TH ^^^ 3 ,3Q ^^^^_ .^j^^^ 5_4g ^_^j_ rj.^^^ ^^g^ j^g^^l^ j^-j^Q^. ^^_ plosions at 11.18 a.m. and 3.33 p.m., while on the east the feeble smoke in the morning suddenly changed to high curdy smoke in the afternoon ; it roared from 7.10 p.m. to 8.40 p.i\[. Both sides were rather active to-day. On the 26th, the west roared at 4.47 a.m. and made a small explosion at 10.17 p.tm. The east was constantly throwing up smoke. Peb 27th— "^^^^ ^^^^ ^^^^ ^^P ash-clouds to a considerable height from morning till 4 p.:\r. and roared 5 times. The weather was misty on the 28th. The east was constantly roaring. The west suddenly woke from a long slumber and rat- tled the city windows, exploded violently at 5.34 p.t^i. with a black column, and roared at 10.53 p.m. TH1-: GREAT ERUPTION OF SAKURA-JIMA IN 1914. 95 Minor explosions at 0.25 a.m. and 0.51 p.m., and a strong roar at 8.39 p.:\i. on ]Marcli 1st. Wet ash-mist covered the whole sou- thern province of Satsuma. The east remained remarkably quiet during the last two days. C) Economic and Social Statistics. People killed. . . .35 (6 in Sakiira-jima) [in l779-'8(). . . . 161— according to some reijort . . . .OjCOO") „ •«•ovinded 127 (1 in Sakiira-jima) ? ] „ fate unknown ... 23. ? ] Damage to jjroperty in the city of Kagi^hima ¥ 1,100,470 ., Saknra-jima „ 3,076,669 ¥ 4,177,139 SAKURA-JIMA. Horses kiUed 1,715 (800 saved) [in 1779-80 1,576 killal] Oxen „ .... 203 (50-GO „ ) [ „ „ „ 135 „ ] Pigs „ .... 612 - ? ] FowLs lost 10,311 ? ] The loss in domestic cattle and fowls amounts to ¥ 52,863 Houses (ménages) in Seto, "Waki, Ari-mura, all on the east side, buried under lava 510 „ „ „ Akamizu, Yokoyama, Koike, all on the west side, buried under lava . . 919 1,429 Houses burnt and buried under lava 1,946 „ remaining which are, however, heii\ily covered with ash 1,186 3,132 The insular population 27,11(> Rescued by launches on Jan. 12th 2,493 „ „ 13th 5 „ ,. 14th 35 „ I „ „ „ 15th 6 2,539 Fled to the nearest mainland bef(jre the catastrophe 24,577 96 ART. 3. — B. KOTÔ: D) Summary and Conclusion.— The writer will now brineriodical]y exploding and throwing up the lava plug solidified within the venthole after the Vulcanian type. The subterranean condition has not yet come to rest. The tremors related in their origin to Sakura-jima and recorded by a seismometer at Kagoshima numbered 667 in Jaimnry, 107 in February, and ;}97 in ^Nlay, 1916. — Nofe durimj Press. THE OREAT ERUPTION OF SAKURA-JIMA IN 1914. 97 nucleus, from telluric or cosmic causes. Vulcanism is the straggling outcome of the processes of general deformation. Since the beginnimi of the present century volcanoes distributed over u'ide regions have been displaying activity. In the South Seas there is a series of volcanoes on the inner side of the fore-deeps of Kermadec and Tonga. To begin from the south, Taupo in New Zealand was active in March 1886, one of the Kermadecs in 1902, Niuafau in the Tonga islands in 1886, Topia in January 1906 and 1907, Fanua-lai in 1906, and lastly, Mt. Matav in Sawaii of the Samoa group in 1902, 1905 and 1906. In this volcanic chain the correctness of the old dogma is again proven, that the alignment of volcanoes is very intimately related to the tectonic structure of the region. Ambrym Island in the New Hebrides, not far from the above-mentioned chain, was the seat of activity from December 6th, 1913 ; Mt. Minnie collapsed on the 12th, and the burning lava welled out, overwhelming all the villages of the north coast. Puna in Hawaii was active on September 21st, 1908 ; Korintji in Sumatra on June 3rd, 1909 ; Sangir in the Dutch Indies on March 14th, 1913. On June 6th, 1912, Katmai in Alaska burst out, throwing up ash-clouds to a height of 12 km., causing an afterglow of the sun in America and Europe. The dust thrown up to the stratosphere was so dense that it had an influence on meteorological phenomena, on the atmospheric temperature ; and many writers came to the conclusion that volcanic dust must have been a factor in the production of past cKmatic changes — the glacial age. Mt. Colima (1901, 1902 and March 24, 1903), Santa Maria (1902) in Central America were active, and Chirique near the Panama (*anal a few years later. In the Atlantic, the eruptions of Martinique and St. A^incent (1902) are well known. Teneriffe erupted in 1909. In the Medi- 98 ART. 3. — B. KOTO : terranean, Vesuvius manifested its activity in 1906 and 1913, and Stromboli in 1907 and 1910. The beginning of the present century seems to mark a general awakening of vulcanism after a period of comparative tranquillity. In the Japanese islands, at the northwest corner of the Pacific, we find also a recurrence of vulcanism. Jensen in his Samoan paper says, that a volcanic island revives its activity after a cycle of sixty-years. In individual volcanoes like Asama, it is likewise after the completion of about 60 years ; but the regional activity of deep origin revives after the double length of the sixty-year cycle. The following may illustrate the case : Oshima Is. in Central Japan erupted in 1777-78^ Saknra-jima in Ivyûshû jj 1779 Aoga-shima | Asama J in Central Japan J5 1780-'85 1783 K 1 Unzen in Kyûsliù >> 1792 The above table shows that one period of the regional activity was completed within 16 years, during which the centres of dis- turbance were regularly and alternately shifted to and fro from Kyûshù to Central Japan, or the reverse, thus illustrating the typical case of a ' sympathy of alternation between two region s.'^^ After a lapse of 120 to 135 years, or about the double length of the sixty-year cycle, the same regions are now being disturbed by vulcanism, viz., Asama, Yaké-daké, Oshima, and the two volcanic islands of the Benin group, i.e., Iwo-jima and Tori-shima (Aug. 1902) in Central Japan ; and the Tori-shima (May 4 [April 11], 1903) of the Eyûkyû chain,"^ Kirishima (Jan. 8, 1914), the Iwô-jima^^ or the ' sulphur island ' (of Kagoshiraa Prefecture, Feb. 1) See page 93, and footnote 3. 2) See pages 4. 3) See ])ages 5, 21 an.l 92. IHE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 99 13. 1914), Suwanosé-jima (Mar. 21, 11)14 to tlie beginning of .luly, 1915, cf. p. 4), and Kuchino-Erabu in southwest Japan. The An-ei ernption (17 70) of Sakura-jima occurred 135 years l>efore the eruption of the same volcano on January r2th, 1914. No one can offer a satisfactory explanation at the present stage of (^ur knowledge about this recurrence. ii. Premonitory Symptoms. — As premonitory symptoms of the paroxysmal eruption, the land at the north of Kagoshima Bay, es- pecially near Yoshimatsu Station, 55 km. from the bay, had been frequently shaken during the eight months before the event. It may be of interest to give here a précis of the ' Report ' by Y. Sataké, Chief of the Meteorological Observator}' of Miyazaki Prefecture. Date, 1913-1914 Number of Earthquakes Period I. pxay 19tli-30th (12 days) 24) Strong eartliq. ] [39 10.39^ [Jnue 2Dd-8th (7 days) 4 strong. . . . 15J / (June 25tli-July 2nd (8 days) 8 strong. . . .31\ a -t Sympatlietic Ijû-in earthq., June 29tli-30th. July 9tli-18th (10 days) 6 strong. . . July 23rd-30th (8 days) 1 strong. . . Aug. 4th-13th (10 days) . . . I'eriod II. (Climax) Str(3ng earthq. 140/ 33 11 . 7 ,37 ,17 136 Period lU. (After-shocks) , Strong earthq. * Aug. 16th-23rd (8 days) 1 strong. Aug. 27tli-Sept. 1 (6 days) 3 strong . No eartbqs. from Sept. 2nd — Oct. 16th (45 days)./ /Oct. 17th— Nov. 16th (30 days) 5 strong 11 \ Tlie first eruption of Klrishiina on Nov. 8th, 11 p. in. No earthqs. from Nov. 17th — Jan. 3rd, 1914 (48 days). The second eruption of KirisMmn on Dec. yth, 4.1ri a.m. Jan. 4th-14tli, 1914 (11 days) .... 3 The third eruption of Kirishimu on Jem. 8th, 2.20 a.m., and lastly, the eruption of Sahura-jima on J'in. 12th, lO.rJ a.m. / 14 100 ART. 3.— B. KOTÙ : To summarize from the data given above, the local earth- quakes at the northwestern foot of the volcano Kirishima began suddenly with moderate intensity which culminated at the second period a, when the sympathetic Iju-in earthquake at the west of Kacroshima shook the ground violently. Even within the narrow O 'CD »/ seismic area of Yoshimatsu the epicentres shifted from west to east. The daily maximal frequency of earthquakes happened 2 hours later than the arrival of the daily maximal atmospheric pressure/^ the average daily maximum pressure of the atmosphere during the earthquake -swarm being 8-12 a.m. and especially 10-12 p.m. ; thus corroborating the long- established rule that high pressure determines the precise moment of quaking, and also of vulcanism. The salient feature noticeable in the above table is, that the volcanic vent burst open during the waning third period of the earthquake-swarm, when shocks became irregular and interrupted with intensified quaking (strong earthquakes 35\). As to the volcanic activity of Kirishima with which the perimetric Yoshimatsu earthquakes had decidedly a close relation, the /zVsi eruption occurred on November 8th, 1913, 11p.m., ejecting ashes and sending up flames, and shattering windows in the city of Kagoshima through air-concussions. The event happened after high atmospheric pressure. The second explosion of high intensity was on December 9th, 4.15 a.m., again after a maximal atmospheric pressure, with ashes and flames as usual. The volcano was treacherously silent till the very moment of outburst, as the writer has often experienced in the explosions of Asama. The third ex- plosion took place on January 8th, 1914, 2.20 a.m., again after maximal atmospheric pressure. Since then Kirishima has remained 1) See the data appended iu the said ' Eeport." THE OREAl^ EllUPTION OF SAKURA-JIMA IN 1914. 101 silent aiul at last, four days later, the memorable eruption at Sakura-jima occurred. The above conclusion concerning the maximal frequency of the earthquakes, which responded to the subterranean activity of Sakura-jima, namely that this frequency, prior to the great vol- canic catastrophe, had some relation with the maximal period of the local atmospheric pressure, is verified by the following facts. According to the ' Report of the Meteorological Observatory in Kagoshima,' the daily frequency of shocks has been high between >>-9 A.M. and 8-0 p.m., corresponding to the pressure-maxima of 8-9, 9-10 A.M., and 8-9, 9-11 p.m. Sakura-jima also burst out on January 12th, 10.5 A.M. after the maximal period of local atmospheric pres- sure. The writer could not find any relation between the direction of pressure gradient and the oè^urrence of shocks. iii. The Phases of Eruptions. — F. Omoei {I.e. p. 57) tabulated his tromometer observations, carried out under his direction in the city of Kagoshima and in Sakura-jima, the registration being started on the 16th, and the writer reproduces it hero in order to explain various phases of activity recorded in his diary (pp. 56-95). The first phase of the Sakura-jima eruptions passed away during Jan. 12th to 13th. The second phase (p. 73) was ushered in on the 14th, continuing till the 20th. On examining the second column, the eruptions of the first order were 60 in daily average from 16th to 20th, these numbers being decidedly the joint records 0Î the eruptions of both the western and the eastern vents. From the 21st the western vents came to a standstill, while the macro- scopic lava movement of the west came to rest only on the 27th. From 21st to 27th — the interval of the third phase (p. 87), the daily average of outbursts was 30 and they were recorded from the eastern vents only, while the macroscopic lava-flowage 102 AfiT. 3.— P.. KOTO : was observable till February 2nd during the fourth declining phase (p. 89) in the eastern lava -field, when the strait of Séto became completely barred. From the beginning of the fourth phase the decrease was very marked. The fifth jjhasc' (p. 91) began with February 3rd, the activity steadily and imperceptibly waning. Outbursts Date 1914 Volcanic earthqs. without eruptions Strong, intratelluric, Weaker, superficial, detonating non-detonating ^.5 fJan. ]6 fl3 ? ce in ^ rr th w in =^ 17 a. 67 1 11 tc * ^ e; IT* ^ ^ IS kl 53 84 20 "H. 5P s rt 'O 19 ^. 74 39 3 §1 'S A\ estern 9 " 20 ft \(35 ^ents came 78 5 U2 »fh ' to rest 77 7 ' 21 37 22 c 29 93 4 t- 23 CD 34 82 2 1 24 40 78 1 r- 25 _>. 36 63 5 2(3 Lava ceased to Ê 29 101 1 flow on W. 27 \35 136 148 4 2 ce / 28 14 ci « 29 17 147 8 a: 30 1 144 3 J 31 5 130 3 "? Feb. 1 6 117 9 1 Lava ceased to flow on E. 2 4 2 108 60 1 2 3 o 4 105 5 .3 '" 5 — 82 0 53 6 2 160 4 7 1 6:i 1 THE (ilJF.AT ERUPTION OF SAKURA-JIMA IN 1914. lOH According to Omoki, the stronger outbursts (second column) were non -detonating and intratelluric, with comparatively large earthquake motions, the double amplitude being 0.1 mm., and they were followed by forcible ejections of ash clouds. The activity, unsuccessful attempts at complete explosion, must have been taking place deep in the channel. The weaker outbursts (third column) were of a surface nature, accompanied by loud cannonading'^ and echoes of sound, and pro- ducing air-concussions which shook windows at a distance. The earthquake motions were, however, feeble, being ten times less than the former. They were the result of the breaking open of the cooled crust of lava vents by pent-up gases. As the detonat- ing velocity was very great, the works of the blasting which were achieved within a very short moment, were correspondingly of considerable magnitude. It is a noteworthy fact that the number A of weaker outbursts, as Omoki says, is inversely proportional to that of the stronger till the end of the fourth phase. The phenomena'^ of eruptions, macroscopically observed daily by the writer on the field, were of two kinds, vu., a) the can- nonading with shocks and quaking of ground occurred at the time when a new or renewed fissure was opened, or a subterranean gush of lava was caused by gaseous pressure conditioned by sudden expansion at the orifices, and b) the thundering and roaring"^ with 1) About the detonations see p. 6G, footnote. 2) The writer carefully observed the night scene of the eastern vents at Fukiiyama. Single booming was accompanied with shocks of the ground, emission of sparks and hurling of red-hot stones. Continuous roarings and clinldngs, which immediately followed the preceding, signalized the outwelhng of fluent lava and the subsequent down- rush of it from the margin of vents. The two did not necessarily follow one another, but occurred quite indeiiendently. After his numerous observations in Stromboli, A. Sieberg found that strong exjilosions begin with weak momentary tremthngs or oscillations of the ground, followed a few seconds later by either a dull long thundering or a sharp booming. At the same time fumarole steam rises accompanied with the trajection of lava fragments. — ' Einführung in die Erdbeben-und Vulkan- kunde Süditaliens,' Jena, 1914, 8. 203. 104 ART. 3. — B. KOÏÛ; little or no trembling (air-concussions), accompanied the breaking open of the cooled ping of the lava vent, followed by the cracking and clinking noise of the overflowing lava-stream. The volcanic earthquakes without eruption (fourth column) showed a marked tendency of standing in opposition to the fre- quency of stronger outbursts (column II). To the writer the numbers of the fourth column seem to have been the movements to and fro of magma in a reservoir more than 1000 ?7i. deep from the vents. Consequently, these ' endo volcanic earthquakes ' have closer relation to the stronger outbursts as compared wâth the weaker surficial activity (column III), tliough the relations are remote with either. iY. Mass and Dimensions of Lava. — The lava-efiusing volcanoes are very rare in modern Japan, in contrast to surficial straggling explo- sive eruptions. The rocks are a basic type of andésites (auganites) of intermediate acidity. In Sakura-jima, lavas are poured out from the Yuno-hira vents on the western slope as w^ell as from the eas- tern, Nabé-yama vents. Both have independent lava fields and both flowed directly into the sea, just like the recent Sawaiian flows. In the following the writer gives the dimensions and the mass of lavas, which are necessarily of the nature of a first ap- proximation to truth, besides he gives the numbers computed by his colleagues for comparison. We find a not inconsiderable dis- crepancy in the columns, owing to the estimation firstly, of thick- ness assumed by different writers after their personal judgment, and secondly, of the extension of submarine flows. The writer is favored with the data of the new soundings^^ of Mr. Fujishiro's 1) The soundings were taken at the end of April, 1914, and the calculations given here are all based on the data at hand at that time. Since then, some changes were brought about especially on the eastern field (Text-fig. 24) by constant movements and secondary flows of the lava-front seawards. The marked changes occurred in regard to areal extent if a comparison is made between the area marked • Marine Lava Field ' from the siirvey of April. 1914, and that outlined with a heavy line, with the remark ' Sept. 1915,' the latter b3ing the result of the renewed land-surveying done by the Prefectural office of Kagoshima. THE (iRKAT ERUPTION OF SAKURA-.TIMA IN 1914. 105 party by Colonel Kanéko, of the Hydrographie Office of the Im- perial Japanese Navy. ,,. The lava-front of the western lava-field (Text-fijj;. LAVA-FIEI.D j,)^ ^ -^^^ ^^^^^ ^^^^^.^ ^^^ j,^^^^^ ^^^^^ .^^Q^^. 24 4 fathoms deep in average, which is eqnal to 44.6 m. As the bottom was gradually and uniformly slanting the mean depth would have been 22.3 m. Since the islet of Karasii-jima (22 m. high), located at the middle of the marine lava-field on the 10 -fathom-line (18.3 m.), was now completely buried (January 19th) under lava, the thickness of the lava -sheet at this point must be not less than 40 m. The thickness of flow on land is also about 40 m. There- fore the average thickness of the entire western lava -field will be approximately 50 m}^ The writer is on this point in perfect accord with all his colleagues mentioned in the Table (p. 107). The field is 4 km. in axial direction and 2 hn. in breadth at the front. A little more basic lava crawled down for a dis- tance of 4 hn. and spread in the same measure over the eastern slope, then finally plunged into the sea and choked up the narrow strait of Séto, 30 to 40 fathoms deep. The lava of this area makes a flat of 120 m. upon the buried slope originally 80???. in altitude, and a hill top on shore, 103 7?2. high (Akashi Gongen, PI. VIII. Figs. 2-3 ; PI. IX. Figs. 2-3), was overflowed with a sheet of lava-flow. The thickness of lava on land may therefore be rekond at 40 m., as in the flow on the opposite side. The contiguous lava-field in the sea has, however, a thickness of 100 m. Eastern Lava-field 1) From experience the writer is disposed to think that the average thickness of a single lava-flow of a femic andesitic magma on land amounts to from thirty to fifty metres, as we can often measure it at the lava-front, as well as in the benches within the steep crater-wall of the Japanese volcanoes. The limitation of the thickness mainly depends upon the viscosity in- herent in the chemical nature of the modern Japanese lavas. 106 ART. 3. — B. KOTO The Southern main arm (Text-figs, 19 [p. 74] and 24) of the iiuent lava pushed southwards under the sea with a thickness of 100 m. where the bottom is now only 30 fathoms, while it w^as formerly 06 fathoms deep, giv- ing a maximal difference of 60 fathoms which is equivalent to a vertical extent of over 100 ?7Z. of the submarine lava- sheet. The eastern submarine arm was deflected east- Fig. 24.— Distribution of the Nabé-yamn lava-flows, land, marine WardS by the head- and submarine. The heavy line, marked ' Sept. 1915,' shows j.^^^^ ^j Sakkabira the subiierial extent of the lava-field by the recent survey (p. 104, footnote). witli Icss intensity of forward movement and corresponding decrease of thickness to about half that of the southern. As will be seen on tlie Sketch Map (Text-tig. 24), the southern arm of the submarine flow moved farther forwards pushing up the muddy bottom to an extent of 28 7/?. This outer muddy shoal is now 60 fathoms deep, while it was formerly 70-80 fathoms. As the soundings brought up nothing but samples of mud, the bottom of the shoal is evidently not sheeted with lava, and the writer deems it better to exempt it from the submarine lava area, althougli in the following calculations it is assumed as a lava mass. THE GEEAT ERUPTION OF SAKURA-JIMA IN 1914. 101 I. "Western lava-field. — Dimensions of lava- streams from the Yunoliira vents. Author Ai-ea Thickness Volume Koto 7.202 sq. km. 40 m. 0.2881 cull. km. A Omori 8.330 40 0.3332 Sato 6.025 40 0.241 Yamasaki 5.23 40 0.2092 H. Eastern lava-field. — Dimensions of lava-streams from the Nabé-yama vents. Aiithor Area Thickness Volume Koto 5.40 sq. km. Lava-field on land ' 1.886 „ „ in sea 2.738 Svibmarine flow on S. ) h 3. 853 1.115 „ „ „ E. j J 4.238 Outer mud shoal 11.139 40 '/*(. 100 100 50 28 0.2160 cub. km.\ 0.188G 0.2738 0.0558 0.1187 0.8529 Omori 5.25 Lava-field on land \ 2.19 „ „ in sea above sea- level [-15.41 7.97 „ ., belou- „ ' 40 100 100 0.21 \ 0.219 !■ 1.226 0.797 ) Sato 6.2875 25 0.1572 Yamasfiki 6.21 25 0.36445 Author The entire lava-field Area Volume Koto 18.341 sq. km. 1.1410 cub. km. Omori 23.74 1.559 Sat5 12.3125 0.3982 Yamasaki 11.44 0.57365 The whole mass 3,012,240 million Kilogrammes or 3,012 milhon tons. 108 ART. 3. — B. KOTO : From tlie Table given the area of the western lava- field is 7.202 sq. km. with a corresponding volume of 0.2881 cub. km. ; that of the eastern is 11.139 sq. km. inclusive of submarine flows, and the volume 0.8529 cub. km. As the density of the Nabé-yama lava is Whole 2.64, the wholc mass of lava poured out both from the Lava castem and western vents amounts to 3,012,240 million kilogrammes or 3,012 million tons. The volume of juvenile lava is equivalent to 1/14 of that of Sakura-jima, the latter being 15.949 cub. km. (p. 32). Assuming that recent and ancient lavas have the same density, the relative proportion of the mass of new lava to that of the preexisting is also 1 : 14. On his trip to Sakura-jima in April 1915, F. Omori took a remarkable photographic view of the lava -flow on the Nabé-yama side (PI. XII. Figs. 1 and 2). This is a secondary lava-stream pressed up fi"om below the still unconsolidated portion of the extensive lava -field of the preceding year. The still molten mass inside l)urst open through the outer hardened crust and crawled downwards from the middle of the field over its gray ash- coated solid sheet toward the lava-precipice. From this point or ' head ' it rushed down anew into the sea (black fresh lower area in Fig. 1), taking a divergent course at its front. The land so newly reclaimed pyrogen etically is, geomorpho- logically speaking, a kind of delta which may be likened to the ' bird-foot delta of the Mississippi.'' See PL XII. Fig. 1. The digiti- form field has ramifying axial channels (black lines in tlie picture), comparable with the diverging ' passes ' of the great American river. While the upper and under portions had solidified and were creeping sluggishly onwards, the still molten portion was able to move faster and thus to leave empty spaces behind it. The steam THE (iÜEAT ERUPTION OE SAKURA-JIMA IN 1914. 109 and gases in the picture are generated in such a tunnel, and gusli out at the end. "J'he channels which present the appearance of natural levees are really collapsed tunnels with broken-in roofs, the vault material being partially carried forwards with the flowing liquid rock. The event is said to have happened some time between the end of March and the beginning of April, 1915, i.e., after a lapse of fourteen months after the great eruption. Another feature, no less striking, is what the writer calls by the name of the gas-blowing horn at the largest tunnel's end, as may be seen in one white spot in the background of PL XII. Fig. 1 (/?). An enlarged and nearer view of it is the lower picture (Fig. 2 (/?)). This volcanic spouting horn is something like that often found in cliff-caves exposed to stormy coasts and like that it is generated in an analogous way, diftering only in mode, which is here by a dry, not a wet process. Pent-up vapor and gases in the lava-tunnel forced their way out, causing this remarkable volcanic blowing horn at the water's edge. It is a special kind of fumaroles. A relinquished lava-grotto is found near Cape Nagasaki (Geologic Map, gr) on the southeastern shore, where it opens at the termination of a lava -tunnel of an ancient flow at the water's edge. Here a cold " infernal " wind, so say the people, gently blows throughout the year from the deep interior. The tromba on land as well as in the sea [postea, p. 115) may be explained, according to the writer's opinion, on the same principle. In the western field (p. 105) the lava flowed down from under a partially consolidated supercrust (on the front of PL XII. Fig. 3) toward the sea, where it deployed into two fan-shaped areas with an indentation between them, the head of which marks the posi- tion of the buried Karasu-jima. The back of the main front ilO ART. 3. — B. kOTÛ : describes a large ring, the edge of which is signaHzed by a series of fumes well seen from a distance. It is the remarkable example of a ring of loet fumaroles around the central arena where dry hot invisible ones exist which prevent the precipitation by the high tension of rising vapor. After rainy weather the fumaroles in- crease considerably in intensity^' — a fact which speaks for the atmospheric origin of the vapor. The arena of this amphitheater was formed by the settling"^ of cooled crust in the unsupported subterranean channel which was made vacant probably through the escape underneath of uncooled magma outwards, and which was unable to bear the superincumbent w^eight of the crust. Lav a -deltas are by no means rare when fluent lava crawls down into the sea ; but what the writer finds particularly remark- able is the wdde circular depression, or ' foundered ' arena, as in Fig. 3, PI. XII., the formation of which must have caused many lava-tunnels in making way for the exit of fl.uent lava into the surrounding sea-bottom. Y. The Temperature of Lavas. — On the intratelluric temperature of the magma within the upper portion of the crust we have only scanty data on which to make even an approximate estimation. There seems, however, not much difference between the initial temperature^^ at the vent and in the channel below 6-8 km. from the surface within the compression shell, as may be deduced 1) See page 116. See also PI. IX. Fig. 1. 2) Pyrogenous settlings «ire teclmically called ' foimclering ' by Daly and other Americau writers. 3) J. P. Iddings also says, that the temperature of the earth's interior is not greater than the hottest volcanic lava, that is to say, it ranges from 1,000° to 1,5()0°C. The soiirce of the hottest volcanic magma, according to this authority and to J. Milne, is located at a dejîth of 30 miles (48 km.), which is the base of the lithosphère at the boundary of the crystalhne and amorphous zones, and the upper limit of the magma region, according to F. v. Wolff, is 30-40 km. Iddings, ' The Problem of Volcanism,' 1915, p. 156 (d scq. "Wolff, ' Dor Vulkanismus,' 1914, Bd. I. S. 31. THE GREAT EBUrTION OF SAKUKA-JIMA IN 1914. Ill from the contact phenomena in limestone which has been influenced b}' injected bathohth of granite, and this temperature is assigned at about 1, 000^-1, 200°C. Wollastonite, a common mineral of contact limestone with granite, also crystallizes below 1,200-. The initial temperature of burning lava at the vent of Sakura- jima must lie above 1130 C. for reasons given below. The melting or solidifying interval of a chemically decom- posable body is rather long, as in the heterogeneous aggregate of minerals which constitute a rock mass. The melting point at the higher phase of the interval depends upon the acidity of the magma. The modern Japan proper lies within the pétrographie province of andésites (Pacific branch)^\ or the rocks of intermediate acidity. As the recent lavas of Sakura-jima have not yet been made the subject of special study, the writer has chosen for comparison the recent lava of Oshima at the entrance of Tokyo Bay. As the physical studies of Japanese lavas are not known in wide circles, and moreover, some results given below are made public for the first time, the writer may be justified in entering somewhat into details, although the lavas under consideration are not those from Sakura-jlma. 'Jhe Messrs. Fuji und Mizoguchi-^ undertook a melting MlHAEA- YAMA p(,)int experiment of the la^a welled out in October 1912 Lava i- i^ from the Miliara crater of Oshima Island. The lava is a liyperstheue-andesite of a rather basic nature \\\Û\ SlOoöl%, Al ,0-^22%, Fe,0,lo%, CaO 11%, and 31(j0 2%. The lava in an electric oven began to be plastic from 800 '-860°, ascertained tln'ough electric con- 1) Koto, ' On the Volcanoes of Japan,' Jour. Geol. Soc. Tokyo, 1916, p. (73). 2) 'The Melting or Solidifying Eange of Temperature of Lava by Electric Conductivity.' Tokyo Sug.-md. Km, Vol. VII. 1914. 112 ART. 'à. — B. KOTO : ductivity, the melting point or the formation of complete melt being 1130°C. At 1300'^ a brisk effervescence or ' storm ' of gases occurred, but not below. With the aid of the optical pyrometer of Holborn-Kuiibaum they found in the field the temperature of the same lava, which in a red-hot state filled up fissures, to be from 995° to lOéS'^C. The craAvling lava stream was examined with the same at night, and it was found that the bright (yellowish-red) j)i"Oximal portion was 995°, the dull (red) terminal poiiion 857°, a temperature at which the rock becomes a little plastic. The writer believes that the lava-streams of Sakura-jiina be- haved, and are still doing so, in the same fashion. Wolfl' (' Der Vulkanismus,' Bd. I. p. 43) says, that foreign andésites begin to be plastic at 1095" and melt completely at 1100^-1125^ Prof. K. Taw AR A, of the Engineering College carried on, at the writer's request, a melting experiment at fixed temperatm'es and for de- finite lengths of time, the action being suddenly interrupted by quench- ing the samples in water at 16°. The high temperatures were determined with a thermoelement. The original rock-specimens Avere collected by Prof. F. Omori from the second period flow of 1912 in the same crater of Osliima island. The lava samples of hypersthene- andésite entrusted to the writer for microscopic analysis are slaggy and black, and dotted with bytownite-phenocrysts. They possess hyalocrystalline texture witli microphenocrysts of hypersthene. The groundmass is built up of feldspar microlites and augite grains, cemented by brown glas:^ Avhich contains dentritic skeletal magnetite. The slides made of the quenched test samples were seen under the microscope with the following characteristics : No. 1. liaised to 900° and kept at this te,inperature for 1 hour. Not nuicli change from the original specimen. No. 2. 900", 2 hours. Dentritic magnetite is still io hv seen. Augite in the groundmass is partially I'used, forming black glass. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 113 No. 3. 1000°, 1 hour. The groundmass-augite is reduced to grains, but minute magnetite ciystals are newly formed. No. 4. 1000°, 2 hours. Phenocrystic plagioclase is resorbed on edges and secondare' black glass-inclosures become abundant. In the groundmass are seen the new augite prismoids and larger magnetite crystals in brown glass. No. 5. 1050°, 1 hour. Phenocrystic hypersthene is rounded a Kttle on edges. Newly formed are roundish rectangular augite and larger magnetite clumps, with colorless glass full of globulites. No. 6. 1050°, 2 hours. Phenocrystic plagioclase is full of negative crystals filled with black glass. A few large grains of magnetite and abundant glpbulitic glass. No. 7. 1100°, 1 hour. The groundmass becomes cr3'staUine, being an admixture of subangular augite and magnetite crystals, with inter- stitial brown globulitic glass. No. 8. 1100, 2 hours. Apparently like the original rock with dentritic skeletal magnetite in glassy groundmass. The test chips are not sufficiently raised to 1130°, which is the complete melting point of the rock. To summarize the feature of changes : the long-prismatic hypers- thene (not amblystegite), which is the only phenocryst of the pyroxene group, remains almost unchanged. The bytownite is left nearly fresh, the only changes noticeable being the increase in size and quantity of black glass inclosures, showing that they are the products of temporary retrogressive resorption phenomena in bytownite rather than the progressive crystal development, as they are usually supposed to he. Feldspar microlites in the ground- mass seem not to have suffered any changes. Feldspar phenocrysts and microlites are usually assumed to crystallize in several genera- tions. To the writer it seems probable that they belong to the same generation, the difference in crystal development being simply due to dissimilarity in their chemical composition. 114 AKT. 3. — B. KOTÔ : The chief changes noticeable within tho range of temperature from 800° to 1100° are those of augite, magnetite and glass of the gronndmass in their relative proportions and also in their development, corresponding to the varying consolidations of lava during the effusive period. The effects of sudden quenching of the test samples under icater do not appear to be strongly marked, excepting the formation of glohullte in interstitial base. Very interesting features to be noted are the duration and the degree of temperature to which the tests are subjected. The longer in time and higher in the scale of temperature which the tests have to undergo, the more perfectly both the augite and mag- netite crystallize from the melts. This possibly depends on the function of time of cooling and the relative proportion of volatile constituents. In connection with the temperature of lava, a few words may ÎPx^^^-i^"^'^^''^ be inserted on tlie formation of the ' block-lava ' and AND BrE4D- BomÏÏ ^li^' ' bread-crust bombs.' The diminution of pressure during the ascent of lava through the channel releases gaseous contents, and the exothermal heat from the reaction of gases counteracts the cooling by expansion and even can raise the tem- perature to a certain degree. In coming to the surface the body of the lava suddenly expands, thereby causing cooling, which is accelerated by the cooler atmosphere. The manner of cooling from outside inward in such masses must result in much mechanical deformation during the forward movement. The projected masses turn to bread- crust bombs and the flowing streams to discrete blocks — the ' Aa ' type of the roughest and most ragged outline. secondaey Four months after the eruption the writer revisited Exhalation ^j^^ terminal cliff of the lava streams of the western THE GllEAÏ ERUPTION OF SAKUEA-JLAIA IN 1914. 115 vents of Sakiira-jima, when bluish wliite fumes of wet disagree- able odor were exhaling from the interstices of the loose pile of block lava (Pis. XIV. and XV.). They gave acid reactions with abun- dant H.,S and a little SO., showing thus the sulphureted-hydrogen fumarolic, or acid solfafaric stage coi'respondhig to a temperature of 100°-200X\ Sulphur w^as being deposited on the walls through the reaction of steam upon H.S. Tests on HCl and SOs gave negative results. The lava-streams flow^ed down on both sides Laya-floav ^ ^^^ directly into the sea, as in the late Sawaiian eruption, calmly generating saline vapor on contact with water, which w^as constantly driven, landwards by convection current, and wdiich enveloped the whole lava- field in dense white clouds (PL VI. Fig. 2, PI. VII. Fig. 1). The lava margin was white with the bubbling up of w^ater. The writer anticipated some damages in orchards and forests from salt vapor^^ as in Samoa ; but the effect was fortunately not pronounced. People frequently spoke of and the writer also noticed that a few metres off the lava front an isolated spiral column of white steam ascended from the sea bottom almost vertically in a narrow thread to a considerable height. It is the wliirlsteam or Ferret's tromba'\ which was constantly changing its position. On land, the same wliirlsteam was frequently seen in tlie unconsolidated lava-field, as in tlie lava-stream from the vent No. 5 of Nabé- yama. 1) This saline steam or Yushhco ('(^ |g), meaning hot saline bath, inflicted considerable damage npon crops on the northwestern shore of Kagoshima Bay during the An-ei eruption os 3779. 2) Zeitsdir. f. Valkanulogie, Bd. I. S. 148, 1914. The writer conjectures the cause of thi- pecuUar gaseous emission to be a forcible escape of comiDressed steam through a narrow orifice at the end of a lava-tunnel, local chamber or the like, perhaps mixed with mist of sea-water See p. 109. PI. XII. Fig. 2 (h). 116 ART. 3. — B. KOTÔ : In passing, it may be noted that white vapor arising from lava-fields both on land and sea (PL IX. Fig. 1) are especially dense after rain, owing to the evaporation of rain water and also to the condensation of saturated atmospheric moisture blown toward the dry and liot interior by convection current.^' Tempera- The tomperaturcs^^ of water on the west at the TUEE OF Sea-watek firgt phase of the lava immersion were reported to be 45° to 50° near the lava front, and 38° at a distance of 100 m. Temperature observations were carried out by a party of the Kyûsliû University at the lava front near the shore at Arimura, on the east, at the end of January, 1914. From the surface to a dejjth of Ô m. thermometers read from 50^-60°, while at the bottom it ranged from 20.5° to 27°, the temperature of the air being constantly at 10°. The writer was on the same spot on April 18th, 1914, and could not approach the lava front witliin a reach of 100 7n., as it was too hot ; but ten metres away from that limit, one could take a warm bath comfortably. The sea water seems to have been heated by direct contact with lava as well as by radiation of subaërial hot lava, which action was confined within the narrow horizontal extent of about 100 m. (PI. VII. Fig. 1) and the shallow layer of a little over 5 m. These ascertained data will be of some service in the study of océanographie subjects. T3 The writer heard froui villagers at Futagawa on ISHES ^1^^ Osumi coast that they saw shoals of ilat lying fishes driven ashore on January lOtli in a^ half Iwiled condition. 1) See page 110. 2) On January 28th, N. Yamasaki (see ante, p. 56, I.e. p. 6) observed at the lava front the surface temperature of water varying from 45.5° to 48°. At a distance of 200 m. it was 30^ (a considerable decrease), at 300 r/).. 21°, and at 600 to 1500 «(., 19.5°. From 2,000 «i. it was normal. In the Sawaiian eruption of 1902, it is said, according to I. Friedlacnder, to have been from 50° to 60° C. THE GREAT ERUPTION OF SAKURA-J[MA IN 1914. 117 becoming half-cooked while emerging from tlic cool sea-bottom through the upper hot zone. The same tiling liappcned in the Sawaiian eruption of 1902. vi. The Lapilli, Ash and Sand. — The recent activity of Sakura- jima was manifested in two ways, beginning with explosions followed by lava efftisions. During the first phase a prodigious amount of disrupted masses was thrown up to the sky, of various size and nature. The western vents were of excessive intensity, projecting blocks of 1 m. or more in diameter, which fell on shore near Yokoyama at a distance of two and a half kilometers from the orifices, and made conical holes in the ground {see p. 66 and PI. V. Fig. 2). The blocks are of juvenile as well as resurgent origin, as will be seen described in the Petkogeaphical Part. We are here mainly concerned with lapilli, sand and ash. According to the usage of mining engineers and especially of agrogeologists various grades are established in the order of size in the skeletal portion of soil in mechanical analysis, viz.. Size expressed in diameter. Name. Minimixm size 10 an Boulder. Varying between 10 mi. .ind 10 mm Pebble. „ „ 10 mm. „ 1 „ Gravel. 1 „ „ 0.1 Sand. 0.1 „ „ U.OI ., Sût. Less than 0.01,, Mud. Kanai,^^ of the Kagoshima Higher Agricultural College, has treated the loose éjecta ù'om Sakura-jima on the sieve method with the following results : a) The ash is composed of hypersthene (not amblystegite), plagioclases and comminuted colorless vesicular glass, of a size less than 0.25 mm., which make up 7 OX to 95% of the whole mass. It is the size which approximately corresponds to that of 1) Journ. (hoi. Soc. Tokyo, July Number, 1914. 118 AET. 3.— B. KOTO : silt mixed with a little fine sand in the grade of the above cate- gory. The particles above 2 mm. are scanty. A wash skeleton (PL XXIII. Figs. 6-7.) of the ash that fell on January 12th on the roof of a house (p. 122) at Kokubu Station, 19 hn. away from the vent, was examined by the writer and seen to be composed of rougMy-rectangular plagioclase, slender hyper sthene and magnetite crystals, the first two being OA^ mm. in longest extension, the last 0.2 mm. The skeletal portion may be taken for fine crystal lapilli or crystal sand in contrast to slag lapilli, as the components remarkably retain their idioraorphic shape. The light major portion consists of finely pulverized colorless splinters of strained pumiceous glass, mixed with small quantities of plagio- clase powder, of a minimum size of 0.002 mm. As the locality is not far away from the orifices, the heavy crystal components are still found in large quantities among the ingredients of the ash. The ash dust, carried away by storm for a distance of 1,000 km., rained in Tokyo on January 14th, two days after the eruption. It is light-gray and is uniform in size, being 0.007 to 0.008 mm. in diameter. It is composed of about ^% of plagioclase fragments, w^hile the rest is glass splinters and textile fibers, with a single piece of hypersthene (PI. XXIII. Fig. 8). The heavy com- ponents— magnetite with sp. gr. 5, hypersthene, 3.3-3.5, and a large portion of plagioclase, 2.5-2.75 — were precipitated during the aerial voyage, and only particles of light friable colorless glass remained in suspension in the air. The glass contains drawn-out air pores, and is marked with striations produced by the stretching of the magma ; thereby the substance became greatly strained and easily pulverulent. It is no easy matter to decide whether the original rock is resurgent or juvenile. From the idiomorphism of the crystal com- THE GEEAT ERUPTION OF SAKUEA-JIMA IN 1914. 119 poncnts it is liighly probable that the asli materials were derived from trituration of pumice. It is therefore to be classed with magma-glass ash, but not lava ash. A question naturally arises : Which vent — the western, Yuno- liira or the eastern, Nabé-yama vent — supplied the major portion of ash materials ? To give a proper answer it is first of all necessary to know tlio date of any ashfall which should have been gathered in a special receptacle, but that precaution was unfortunately not carried out. As will be immediately stated in speaking about lapilli, the main ash producers were the eastern vents {see ante, p. 68, footnote), for the reason that the magnetite found possesses well-defined crystal form and glass pure white, while the western vents hurled out only dirty glass particles and magnetite clumps. h) The SAND {moyezuna or ' burning sand ') varies in size from 1 to 2 mm., corresponding to the gravel in the above scale, though the one witli 3 mm. is tolerably abundant. Like lapiUi, it is really rock particles, differing from them, however, in its com- pact texture, the color being usually black. Undoubtedly it must have precipitated in a large quantity at the very beginning of the eruption as a rain of sand of old lava, when vents first broke open through a preexisting lava sheet of volcanic body. It has contributed in large measure to the darkness of the cauliflowers^^ (PI. I.) of dust clouds, which gave an awful impression to persons who happened to view the scene from far and near. Towards the climax of paroxysm the sand became rarer. 1) The form produced is similar in principle to the smoke ring of tobacco-smokers. The puffed vip vapor in the air from a volcanic vent expands in the horizontally rotating axis in shaping that curdy salmiacal ring, and the solid particles intermixed in smoke columns are sorted and fall leeward as a black rain of volcanic ash and sand. See A. Sieberg, ' Einführung in die Erdbeben- und VulkankiTnde Süditaliens,' Jena, 1904, S. 203. 120 ART. 3.— B. KOTü : At the later declining phase it again increased in proportion as the lava cooled and acquh^ed viscosity, thereby affording great obstacles to outbursts which, when carried to effect, caused scatter- ing and spreading of the rain of sand, by this time of juvenile lava. The writer met with such rain on his second trip. The sands and ashes are therefore of resurgent {lava ash) as well as of juvenile {magma ash) origin, so that chemical analysis made of them seems to be of restricted value, except for agricultural chemists. If samples are gathered at special periods the results attained are quite different. c) The LAPiLLi^^ (' karu-ishl ' or ' ga-ishi ') are fragments of gray pumiceous lava of a size of 2 to 4 mm,, making up from 7 to d7% of the mass. The rest varies in size from 2 to 4 cm. corresponding to the fine pebble in the above scale. They are no doubt of juvenile origin formed during the first phase in the Strom- bolian stage, and should be classed with the slag lapilli in contrast to the crystal lapiUi."^ The writer is able to discriminate the eastern, Nabé-yama, from the western, Yunohira, lapilli. The eastern are in appearance whitish and fine vesicular with roundish outline. Microscopically, we find idiomorphic olivine and hypersthene crystals, grouping themselves in local patches, and magnetite crystals as well as plagioclases are well crystalHzed in outline. Minute roundish air pores (0.07 mm.) are enclosed in the highly strained colorless pure glass. In short, the eastern lapuli are the froth of lava. The western lapilli, on the other hand, are dirty brown and coarse with stretched air pores. They are characterized by the 1) See the heading ' Recent Lapilli ' in Pétrographie Part. 2) Idiomorphic anorthite crystals from Myake-jima and the Volcano Tarumai are well- known examples. THE GREAT ERUPTION OF SAKUEA-JIMA IN 1914. 121 absence of olivine and the anliedral form of magnetite. Tlie glass base is full of dust or a little brownish, and contains a few microlites of both augite and plagioclase. The whole structure appears not unlike a spongy mass with skeletons, enclosing elongated pores of more than 0.3 mm. in diameter. As regards quantity, the lapilli from the eastern vents are far more numerous than those from the western, chemically the former are rather basic, the reason of which is the presence of a basic mineral, olivine. Chemical Among cjccta after a volcanic activity the product Composition ^-^^^ j^^g j-^^^^^ usually subjcctcd to chcmical aualysos is ash. So far as the writer is aware, no definite conclusions have yet been drawn from numerous chemical treatments, owing to the heterogeneous origin and composition as well as varying proportions of mineralogical ingredients dependent upon the dis- tance fi^om vents. In collecting samples fresh ash must be chosen with mention of date and locality, which is usually neglected. A few specimens of analyses made of the recent Sakura-jima ash are given below : 122 ART. 3. — B. KOTÔ : Asli which fell at the Normal School of Kagoshima city Unknown locality- Ash which fell in Krakatoa, 1883 Imp. Geol. Surv. Japan Engineering College, Kyûshii Imp. Univ. Winkler-) SiO, ALOs Fe,0, FeO MgO CaO Na,0 K,0 H,0 TiO, P.O, MnO Cl 63.39'> 16.75 3.10 4.10 1.43 5.38 3.50 1.32 0.63 0.57 0.20 0.03 60.38 18.10 8.49 3.68 6.01 0.56 0.46 0.14 0.07 61.ß6 17.77 4.39 1.71 2.32 3.45 4.98 2.51 1.12 0.41 All three samples of ash derived from hypersthene-andesite, fell near vents, and there exists general agreement among tliem. A soluble portion of ash that fell on January 12 th at Kokubu Station (p. 118) was tested by Mr. limori under the direction of Prof. Ikeda, of the Tokyo Imperial University, with the following results : CaCl, 0.049%", NaCl 0.015%, sulphates trace. It may be remarked that at the first phase the ashes contained clilorides while sulphates occurred only in trace. NaCl is of primary volcanic origin, as the writer took the sample before the inflow of tlio lava into the sea. The acidity^^ lessens with time as shov/n in the following, the numbers given beinor reduced to the standard after Dr. Daikuhara : 1) 15.75^",; of soluble silica. 2) ' Report of the Krakatoa Committee of the Eoyal Society.' London, 1888. 3) Daiknhara informed the writer, that the acidity exists in the soluble portion and is mainly due to the presence of AlCl^. It is Chloraluminite with the composition ot AlCl^.&H^O, and crystallizes in hrxagonal-rhombohedric. Wolff, ' Der Viilkanismus.' Bd. I. THE C.EEAT ERUPTION OF SAKURA-JIMA IN 1914. 123 Locality Date, 1914 Jammry 25 April 8 July 6 Middle July West Kokubu, Aira district. Take in Futagawa, Kimotsuké district 19.1 18.2 1.6 2.1 trace : The much talked- of injurious action of juvenile soil upon plantations will soon disappear after a few months, by rainfall, through the leaching of the soluble salts into the fast soaking bottom. Southern Kyushu is a plateau of Diluvial pumice bed, more than 200 m. thick, formed by deposition of éjecta of ancient volcanoes. Eecent deposits are exactly of the same kind as those of past geologic ages, they are all able to produce fertile soil and are w^ell fitted for pasturage and certain kinds of subtropical plantations, as in the case of Central America. The only defect is the dearth of water and on this account the land could not well be brought under cultivation for rice paddies. ' The mode of deposition of eiecta is interesting and Deposition ^ j. j o OF Ejecta regular. It cau be best seen at Kurokami at the foot of Nabé-yama, where the maximum tliick- ness attained 1.8 ???., burying houses even to the eaves (Text-figs. 22 and 25). Banks of a dry river nearby expose a good profile. a) The basal de- ix)sit is an accumula- tion of coarse angular fragments of preëxist- Fig. 25. AsL-buried village of Kurokami on the eastern shore. 124 ART. 3. — B. koto: ing black compact lava mixed with juvenile lava, superimposed by h) a coarse admixture of lapilli and sand. Near the upper horizon the writer, to his surprise, picked up white porcelain-like splinter blocks, which under close examination were found to be the in- teresting cordierite-bearing éjecta to which special description will be given under the new name of ceramicite. It is remarkable that it constantly recurs in the same horizon elsewhere, c) The upper 13 cm. of the deposit is composed of ash mixed with little sand. The above profile tells the order of volcanic ejections begin- ning with coarse and closing with fine materials. The lower portion of the (a) deposit is of projectiles during the explosive phase, while all the r(^st are the products of the eruptive period. The new deposit affords good ground and is pleasant to walk over, being a typically macadamized terrene covered with moist fine ash which coats hills, mountains and valleys alike. What is said in the foregoing chiefly refers to the normal deposit. On the west, however, abnormal conditions prevailed. Here the first paroxysm seems to have been an extraordinary one, scattering rough blocks of 1.2 m. diameter, of old compact and new pumiceous lavas on the shore between Akamizu and Akobara, at a distance of 2 km. fi'om the vents. The loose pile of about l.Zm. was coated with 1.3 c??z. of ash sheet. The ground here is of roughest character and very unsteady (PL V. Fig. 2) in contrast to the east (Text-fig. 25). DispEBsAL ^^^^ mode of dispersal of the éjecta was mainly in- fluenced by the direction of the wind, especially the coarser particles, which were under the direct control of varying local winds. A photographic picture taken a few minutes after the first eruption (Jan. 12tli, 10.20 a.m.) unequivocally shows the wind veering at that time to northwest (Frontispiece and Text-fig. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 125 15, p. 64), blowing coarse products as well as fine dust to the northwest, north and northeast, along the foot of the main volcanic cone, wliich locally deflected the course of projectiles to a northerly direction. Coarse éjecta are therefore remarkably absent on the southwestern coast of Sakura-jima. Pig. 26. — Distribution of éjecta in Southern Kyushu. Main volcanic ejections occurred from the first moment of subaërial activity till 4 a.isi. of the next day, the 13tli. Meanwhile the wind changed to W.N.W., driving the showers of lapilli to the province of Osumi [see Sketch-map, Text-fig. 26). The fine dust was thrown up to a height of 18 km., as was estimated by the Station-master of Kagoshima, though this stands far below that of tlie Krakatoa eruption, which was from 27 to 29 km. {see 126 _ ART. 3. — B. KOTÔ: p. 65), and the explosion gases ascended up to 70 to 80 hn. As it was above the lower limit of the zone (11 A'??î.) oj windless constant temperature (-GO°C), the Sakura-jima dust may have remained in suspension for a long time. It is here to be remarked that the height calculated from Ku- shino, far away west of the city is 7,272 m., that estimated from the war ship ' Tone ' anchored near Taniyama is 8,181 m. The volcanic gases, which accompanied the eruption, must have ascended still higher and pressed up the zone of constant temperature to a cooler region, causing copious precipitations of rain. The writer has fre- quently asserted that the cloud-burst, however, did not occur with a rain of mud ; on the contrary the first phase of eruption happened in fine weather. Possibly the recent eruption of Sakura-jima was more or less of the anhydrous character {see p. 92). ^y^'^ ^^ Major portions of fine éjecta floated in the higlier "mEßl' atmosphere and were transported eastwards by the constant upper current of a westerly wind. Ash fell in the four provinces of Satsuma, Osumi, Hyûga and Higo in soutliern Kyushu. (Text-fig. 26). At Miyazaki (80 hn.) in Hyûga it rained from 1.45 a.m. of the 13tli. The whole Chiigoku region of Honshu was hazy on the same day. The island of Shikoku, especially the southern half, was the heavy precipitation area. In Osaka, dust fell slightly from noon of the 13tli till tlie morning of the 14tli. Thence eastwards it became thinner, and in Tokyo, at a distance of 1,000 hn. from Sakura-jima, it fell from the morning of the 14tli. The same thing happened in the eruption of 1779. It is not surprising that dust fell in Tokyo ; for, the Icelandic eruption (Rudloff crater) of 1875 precipitated dust at Stockholm after 30 hours at a distance of 250 miles (1,900 A;??!.), and the Krakatoa eruption at a distance of 2,500 hn. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 127 In the Bonin islands an afterglow of tlie sun, as in the Krakatoa eruption of 1883, was observed and dust began to rain from 5.30 a.m. of the 13tli. It was transported thither with a velocity of 30 m. per second. Some doubts were entertained as to the source of the Bonin ashfall, as a new volcanic island rose on January 23rd from sea-level near South Iwô-jima, which lies to the south of the above-mentioned group. The wi'iter abstains from calculating the amount of éjecta from the Sakura-jima vents, because of the want of data for trustworthy estimation in sucli loose volcanic products of various dimensions.^^ A According to F. Omoei, the volume is 0.62 cub. km., which is equivalent to 4:0% of that of new lava, and 1/25 of the volcano of Sakura-jima above sea-level. The ratio of the volume of éjecta to new solid lava is 8:5. In the Krakatoa eruption of 1883 (loc. cit., p. 122, footnote 2) the total volume of pumice and dust was estimated at 9^% of the whole éjecta, there being no lava flows at that time.'^ The fine pumice of that event was roughly calculated to be 3.5 times and the coarser sort 5.5 times lighter than the compact lava. Yii. The Subsidence of the Sakura-jima Environs. — After the cat- astrophe of Sakura-jima on January 12th, 1914, it was rumored that the neighborhood was gradually subsiding. Unequivocal signs were at hand : the overflow of bay waters into paddy fields and salt gardens over and through embankments on shore, thereby endangering both the rice culture and industry. Houses at Futa- mata and Shirahama standing on the northern shore of the volcano island became exposed to spray upon their w^ndow^s, where they 1) In Kiigoshima it was estimated that the dry ash occupying a sq. m. with a thickness of 2 mm. weighed 118 grammes. 2) In the CotopaxL eruption of 1880, the quantity of ash was estimated as 2,000,000 tons ; that of Katmai occupied 5 cub. miles. 128 ART. 3.-B. KOTO: were formerly outside the reach of rollers, and in April the writer frequently overheard people talking about the land settling 2 or 3 feet. The same thing happened after the eruption of 1779. It is stated'^ that the sea encroached upon the city of Kagoshima and its northern extension of shore by a subsidence of 5 or 6 and even 10 1 feet. The immerged land seems to the writer to have been later reclaimed, though some recent writers entertain the view that the lost land resumed its former altitude after a score of years. The bottom of the trench bay, or geologically speaking the rift valley, of Kagoshima is settling even at present, as will be immediately spoken of. As to the recent positive movement of sea level, a popular view is that the current is now being deflected and stemmed to the opposite shore through the complete damming up of the strait of Séto by the new lava-flows (p. 90 and Text-fig. 19 in p. 74), wliile a few assign the cause to the increment of water tempera- ture through the inflow of ' live ' lava into the sea. Thanks to the authority-^ of the Ordnance Department of The Headquarter Staff of our Army, the writer was given all the available data of former and recent levelings undertaken with the express purpose of ascertaining the changes, if any, of levels in the neighborhood of the volcano after the paroxysmal activity. The resurvey resulted in the affirmative, and more than that, the local disturbance was found to be confined within a circular tract having a theoretical radius of 61 hn., conforming to that post- 1) An episode in Mr. Nankei Tachibanu's ' Seiyiiki ' or a Journey throur/h Soïdhern Japan, a ne-ws item in the Katjoshhmi, Daily, of October I8tb, 1914. 2) The resxilt was later piiblisherl in a pamphlet with the title, 'The Cmst-shifting and Relief Changes in Kagoshima Prefecture after the Eruption of Sakura-jima,' May, 1915, Tokyo. The writer received it only at the end of 1915. THE GREAT ERUPTION OF SAKÜRA-JIMA IN 1914. 121) ulated by G. K. Gilbert*^ to show the quantitative relation between local variation of depth of compensation and the resulting variation of gravity. The nature of the disturbance in the present case is expressed in the relief through the concentric settling of terrene in a cauldron -shape with a centre off the northern shore of Sakura- jiraa (PI. XIII.). The subsidence of the present area involves a problem of isostasy, caused directly by a mass defect produced in the out- pouring of lava of 3,012,240 million kilogrammes or 3,012 million metric tons (p. 108), besides an indefinite quantity of éjecta hurled up from the vents and distributed within a distance of more than 1,200 km. (to the northeast of Tokyo). H. Nagaoka, of the International Geodetic Committee, has made known in a series of Annual Reports the anomalies of gravity at Oita, Nobéoka, and Myazaki on the eastern coast, and at Shibushi, Kagoshima, and Hitoyoshi oa the south, and in the interior of southern Kyûshù, the former 3 with negative and the latter 3 with positive values. These sets'"^ of values show that the region has been, in general, out of isostatic equilibrium, and this has been accentuated by the recent efiusion of lava at Sakura-jima, w^hich lies, so it seems to the writer, on the borderland of positive and negative anomahes. The borderland is as a general rule apt to be subject to unbalanced stress accumulation, as in Central Japan, where the great earthquake of 1891 devastated the stretch between the t0^vns of negative Gifu and positive Nagoya. Assuming that the zone of compensation lies below 122 or 1) •Interpretation of Anomalies of Gravity.' Prof. Paper 85 C. U. S. Geol. Sun:, 1913. 2) Oita î7"o-7'o -0.032| ShibiisM g'\,-y'„ +0.034 Nobéoka „ - 0.008 [ ^^^* Kagoshima „ +0.045 I const Myazaki „ -0.029' Hitoyoshi „ +0.000 J 'Ergebnisse der relativen Schweren messungen in Japan in den Jahren 1909-1911 und 1912.' South and inland 130 AKT. 3. — B. KOTÔ : 113.7 A-??2., Gilbert attributes the cause of gravity anomaly to varia- tions in the vertical distribution of density. A local, but not re- gional, anomaly of defect, with which we are now concerned, means a defect above and excess below, tlie sum of both being the same everywhere at 122 hn. The inequality adjusts itself by undertow or underflow in the zone of relative mobility which can originate in liquefaction near the surface as a consequence of reduction of density through faulting, and the mobility of the in- tercrustal layer further causes vulcanism.^^ A regional anomaly defect of mass is, however, supposed to be adjusted by undertow above, below or within the zone of compensation. According to Barrel"^ the undertow is only possible below the zone in the asthenosphere or the zone of weakness. Having traced the relations among undertow, mobile zone, diastrophism and finally vulcanism, we now turn to our subject proper. ^"equL""^ In the annexed Sketch-map, PI. XIII., the ' curves Subsidence ^£ equal subsidcuce ' worc drawu approximately through the points of equal settling of terrene.^^ To each point are attached the figures in millimetres reduced to the level of Tokyo, which express the differences between the levelings done during 1892 to 1900 and the resurvey during July and August of 1914. The writer has drawn ten curves, viz., of 800, 700, 600, 500, 400, 300, 200, 150, 100, and 0 mm. The curves describe roughly concentric circles which are at first closely packed together near Cape Osaki at the north of Kagoshima, and there lies a maxnnum 1) Volcanoes usually make their appearance in the region of mass defect rather than in the compacted area. 2) Jour. Geol. Chicago, 1914. 3) A similar attempt has already been made by F. Omori in his recent papers. See Bull. Imp. Eiirthq. Com., Vol. \H1. No. 1, p. 29, and • The Tôyû-gakugei-Zasshi ' {The Oriental Science), :No. 402, March, 1915. 'JHK GKEAT ERUPTION OF SAKUKA-JIMA IN 1914. 131 land depression of 804 mm., wliile outer circles distend considerably westwards. The region embraces the plateau of Yoshino, 500 m. high. (3n the opposite eastern shore of the Bay of Kagoshiraa tlie circles are nearly equidistant, though mostly of schematic nature. , The outermost extension of disturbance from an hypothetic centre lies at a distance of 52 km. on the east, where the settling of level sinks to minimum with a negative value of -11 mm. which is within tlie bounds of observational error due to refraction. (Jut- side of the above border, the values are rather high from unknown causes, though they run below -20 mm. The corresponding western limit cannot be fixed owing to the want of data at disposal ; but the line probably passes beyond the land. From the above, we see that the outer limit of disturbance is not far away from the postulated boundary of local gravity anomaly assigned by G. K. Gilbert, that is to say 61 km. As already stated elsewhere (pp. 57 and 99), the premonitory indications of recent vulcanism were signalized nine months before in a swarm of earthquakes in the Yoshimatsu region (55 km. distant) and the unusual activity of the neighboring Kirishima volcanoes. The hypocentre was later shifted southwards to Ijii-in, and at last the subterranean commotion was satisfied in finding vents in Sakura-jima. All the recent local subterranean agitations confine themselves within a boundary of about the postulated 61 km. from an hypothetic centre of depression in the Bay of Kagbshima. See PI. XIII. Mixo-owAKi jj^ ^jj^^g connection, a few words may be said of Eakth- -^ '^'i^; "^ the gi'cat earthquake of Central Japan in 1891. The result of renewed levelings by The Headquarter Staff' soon after the event showed the radius of disturbance in the Mino-Owari 132 ART. 3. B. KOTiJ : piain to be 70 km., while in the present area it is about 52 hn., the maximal positive and negative w^arpings in 1891 being 767 and 308 mm. respectively. Perhaps these values of radii indicate the limit of sustainment of local cowpetent structurG of the earth's supercrust. Abnormal cases of the curves of equal subsidence in the present area are not rare. The flat- topped headland of Sakkabira, 325 ?w. liigh, facing the buried strait of Séto, is crowded with 3 curves (-700, -600, and -500 mm.) of high value (s:^^'e»"«i and radial fractures of a volcano hai^pcned during, or as a cause of the great Hawaiian earthquakes of 18G8. Evidences collected by H. O. Wood seem to show that they were tectonic, rather than volcanic, earth- <]uakes. That the disturbed area was about 375,000 sq. miles, points to the fact that the depth of origin must have been considerable, and that it gave rise to important oceanic sea waves. JhdI. Sels. Soc. Avierlm, Vol. ÏV. 1914. Tn passing, it may bo mentioned that the destructive volcanic earthqiiakes of Sakura-jima on January 12tli, 1914, were recorded in Tokyo (1,000 km.), and in Laibach in Austria. Xaturt, 1914. The writer has been long since inclined to think that the volcanic and tectonic earth- quakes are by no means easy to separate. We must now take into consideration the crypto- volcanic earthquakes which may eventually cause subaërial volcanic manifestations. TRI': (HiEAT ERUPTION OF SAKUKA-.IIMA IN 1914. 133 \vith the high negative value at Cape Osaki (-894), there Kes amidst the ]^>ay of Kagoshima an hypothetical and geometrical centre (marked with a cross, x, in PL XIII.) around wliich the lines of equal subsidence describe concentric circles with a radius of 52 km. The point is presumably the maximal locus of sub- sidence, which can only be proved by later soundings. Cause op ^j^i^q ]^qI^ mobilc layer or 0. Ampfeee;e's under- SUBSIDENCE -^ current^ ^ was forced up under hydrostatic pressure through the vents in Sakura-jima in the form of ' live ' lava, and the cold heavy crust sunk down to compensate tJie mass defect, which produced on the surface the effect in the form of a kettle- shaped depression. That the negative movement of terrene has not yet come to a standstill, but is still progressing downwards, can be unequi- vocally proved by the differences of values fi'om the first and the second recent levelings. The first was performed during July and August, 1914, and the second during February, 1915. From the values of the second levelings given on bench marks {see PL XIIL), we lind where the high values obtained in the first are cor- respondingly high in the second. The maximum depression'^ of bay shore at Cape Osaki with -894 mm. is now augmented by -O^jo mm. during an interval of 215 days, the hypothetic centre still remaining in the same locus. Should the settling continue in the same fashion in the near 1) ' Ueber uns Bewegnngsbild von Faltengebirge.'. Jahrb. d. k. k. Geol. lltlchsanstalt, \o\. 5r., 1900. Modern views on the formation of mountains based on the Pratt-Hayford principle of isostasy are at present widely discussed in various scdentific publications, and the general ten- dency on this line of study is concisely depicted in K. Andrée's 'Ueber die Bedingungen der Gebirgsbildung,' Berlin, 1!J14. 2) For instance, -(055) expressed in the term of iraiK affixed to the former value -894 )/u/-. at Cape Ôsald. See PI. XIII. 134 ART. 3. — E. KOTü : future, the outstanding plateau of Yoshino will be icarpcd up and cloum in consequence of the overthrust of the surface crust toward the ever sinking centre of depression. So far we have l)een concerned with the vertical elements of crustal movement, we now turn to the horizontal. Comparisons of longitude and latitude measurement of bench marks disclose a considerable discrepancy in the values of earlier and recent dates. Assuming the triangulation base. Order II. A-G in Map, PI. XIIl. as permanent and unaltered, the new geographic positions of bench marks are calculated from actual observations in relation to tht^ neighboring fixed triangulation base, as in Table. No. X Y Resultant Direction of Shifting nil an cm 1 -fO.44 + 0.42 0.61 N. 44° E. 2 + 0.27 + 0.48 0.55 N. 61° E. 3 -f0.05 -0.02 0.05 N. 21°W. 4 + 0.16 + 0.63 0.65 N. 75° E. 5 + 0.34 + 0.89 0.95 N. 69° E. 6 + 0.06 + 0.36 0.36 N. 80° E. 7 -0.14 + 0.79 0.80 8. 80° E. 8 -0.08 + 0.27 0.28 S. 74° E. 9 -0.52 -0.09 0.53 S. 10°W. 10 -0.10 -0.09 0.13 S. 42°W. 11 + 0.02 + 0.01 0.02 N. 25° E. 12 + 0.06 + 0.12 0.13 N. 64° E. 13 + 0.30 + 0.15 0.33 N. 27° E. 14 -0.02 + 0.14 0.14 kS. 82° E. 15 + 0.06 + 0.32 0.32 N. 80° E. 16 + 0.12 + 0.18 0.21 N. 56°E. 17 -3.07 -0.08 3.07 S. 2°W. 18 -3.48 -0.97 3.62 8. 15° W. 19 -1.88 -0.80 2.03 S. 23°W. 20 + 1.08 -0.06 l.ds N. 4°W. 21 + 4.49 + 0.56 5.55 N. 7°E. 22 + 2.37 + 1.11 2.61 N. 25° E. 'J'he X signifies meridian direction, the Y that of parallels. Kos. 1-22, affixed on the ujjper left-hand of the triangles in I'l. XIII. mark the position.s and the nimibers of the bench marks. THE GKEAT ERUPTION OF SAKÜRÄ.-JIMA IN 1914. 135 From the new values of longitude and latitude, X and Y, the resultants are gi'aphically estimated which express the directions and relative measure of horizontal shifting of ground on the posi- tion of bench marks, as shown with arrowheads on Map, PI. XIII. An inspection of arrowheads shows that they also point at the hypothetic centre of subsidence within the Bay of Kagosliima, only Nos. 17, 18 and 10 on the southwest slope^^ of Sakura-jima, however, point at the opposite direction, and those on the high flat of Sakkabira on the southeast are directed toward a secondary centre near the bench mark No. 12. The writer cannot give any adequate explanation on these apparent anomalies. In summaritinçi what is stated above, our Headquarter Staft' did good service in the cause of science in laboriously working out the actual occurrence of both the vertical negative movement of the , crust and the horizontal shifting of geographic elements around and in Sakura-jima after its paroxysmal eruption. A novel feature on vulcanology was brought to hght in the concentric circle of equal subsidence and the horizontal shifting of ground toward a common centre in the bottom of the Bay of Kagoshima, where the maximum subsidence of the crust is to be expected. These movements may well be explained as a local disturbance of the lithosphère on the principle of isostasy. Gilbert (p. 12!)) postulated the radius of local disturbance of gravity anomalies to be 01 /.???., and these anomalies arise from irregular vertical distribution of density with a defect above and excess below, all above the zone of compensation at 122 hn. The in- equality is adjusted by the undertow of a shallow layer made mobile 1) Ihe abnormal shifting of bench marks seems to tho writer to be due to the tilting and slipping of a portion of mountain mass through the shocks given by volcanic spasms. The same explanation can be applied to the bench mark No. 21. See p. 132. 136 ART. 3. — B. KOTÔ: by reduction of density as a consequence of crustal dislocation. The undercurrent so started finally causes vulcanism. The limit of the subsiding area in the present case is at a radial distance of 52 km., while in the Mino-Owari plain at the time of the earthquake of 1891 it was 70 km., approaching in both cases the postulated limit of local disturbance of gravity anomalies. The premonitory wandering earthquakes of Yoshimatsu and Ijû-in, and the unwonted activity of the Kirishima volcanoes for nine months prior to the Sakura-jima eruption all happened in the locally dis- turbed area within the prescribed radius, {sse Premonifory Symp- toms, p. 99). A few points escaped the notice of the writers on the recent eruption, which ought not to be passed over without comment. First of all, it is to be remarked that during the destructive earthquake on January 12th, 6,39 p.m., the first day of eruption, a A solid basalt cliif, 400 ???. high, collapsed near Cape Osaki (sse Map, PI. XIII.), where the railway track and telegraph posts were heavily damaged. All communication with the outer world was then cut off and caused a rumor to spread all round the globe that the city of Kagoshima was buried under lava (p. 70). It seems evident to tlie writer that the accident was not merely due to gravity, but the actual tossing of the subsiding ground induced the heavy slip of the chif. As it will be remembered, Osaki is tlie spot of maximal subsidence on the land side, amounting to 894 mm. Simultaneously with the landslump, the village Game, 12 J>m. northwest from Cape Osaki, was strongly shaken, causing collapse of bridges and cracking of house walls, besides gushing of ground- water from fissures many cho (1 clio=:109 7îî) in length. Cape Osaki is an edge of the Yoshino plateau, and Camô lies on the THE GREAT ]:RUrTION OF SAKUBA-JIMA IN 1914. 137 nortlicni foot of the latter. The region as a whole is unstable and settling at a rapid rate. Lastly, Mr. Kazuno, the efficient director of the meteorological observatory of Kagoshima, warned people up to the very moment of eruption that the epicentre of the incessant earthquakes was in the above-mentioned flat of Yoshino (Geologic Map). This the seismometer indicated to liim and he kept persistently and boldly to his assertion, though the public would not believe him. Finally, the excited people of the afflicted Kagoshima region spread malignant gossip about liim without cause. As will be seen on the Geologic Map, the Yoshino plateau is an uncompensated area constantly sinking down to adjust local equilibrium. The crypto- volcanic earthquakes of the Sakura-jima eruption may have strongly agitated the Yoshino plateau, as his reliable seismometer led him to believe. We are now enabled to approach a step further in tracing the cause and effect of the local isostatic vulcanism, displayed at Sakura- jima in the escape of gases and extravasation of lava of 3,012 mihion tons and an indefinite quantity of éjecta, in the geologic rift-valley bay of Kagosliima. YÜi. The Yentholes.— .VU the scientists, home as well as foreign, who were in tlic field and studied the nature of the district, came to the inevitable conclusion that the lava effusion of 1779 on the southwestern and northeastern flanks of Sakura-jima occurred along the fracture that runs through from Kirishima to the volcanic chain of Ryûkyû, and the recent vents lie on a line wliich cuts the former at right-angles. (See Geologic Map and Text-fig. 10 c.) What is stated above well agrees with the current dogma that volcanoes sit upon the intersecting point of tectonic lines, as we are ac- customed to find delineated on maps of Ischia, Krakatoa, and Hawaii. 138 ART. 3. — -B. KOTü : The writer was probably tbe first amongst specialists, if he may be called so, who hurried to the scene and viewed the front or west side of the grand volcanic display, without be- ing able to come to any definite idea as to the i-elation of vents and geologic structure ; but on seeing the lineal arrangement (Text-fig. 19 ; PL XL Figs. 1-2) of bocche on the rear, the Nabé- yama side, the writer immediately conceived the existence of a mountain fracture extending on both flanks of the volcano. During preliminary work to learn the distribution of the various lava flows, both ancient and historic, and also the history of the building-up of the volcano of Sakura-jima, the writer's view, however, became somewhat modified. Though apparently a simple overtowering konide, Sakura-jima is really a triple one, piled on the shoulder one after the other in a meridional direction in the order from the north top, then the south, and finally the middle (Text-fig. 8, and PL III. Fig. 2). The constructive mature age of vulcanicity having passed away, the declining period of pericentric flank eruptions was ushered in in historic times, when lava poured forth around and outside of top-craters all round the slope after the fashion of the petals of a flower, firmly sheeting the old composite konide with a lava coat of mail. There are, however, portions of the slope still remaining free of lava flows (Geologic Map), and these intervening gaps — tJ/e j^ositions of comparatire weakness in the superstriiciure of the volcano — icere selected for the effusion of hot magma in the recent eruption. More- over, the positions of vents lie at the overlapping junction-line of the north and the south cones which are the dominant topographic elements of the composite triple konide of Sakm-a-jima. The préexistence of a transverse fracture is universally ad- THE GREAT EUUFITON OF SAKURA-JIMA IN 1914. 139 mitted and cited on tlic fact of the presence of tlie parasitic volcanoes of the Nabé-yama homatc^ on the east, and the Hikino- liira dome and the Hakania-goslii patch, besides the lava-drowned Karasu-jima on the west, on the supposed tectonic line (Text-fig. 10.) This argument is not at all conviiicinr/ to the writer, as the four masses difier in age nnd are heterogeneous in their origin. The writer admits the fact that the first is a true parasitic cone ; but the nature of the second, — the Hikino-hira monadnock, is not yet thoroughly established beyond doubt. It may be a volcanic plug or a remnant of an edge of old disrupted slope, the rock being identical with that of the north cone. The third, — the Hakama- goslii patch, is a geologic block of sedimentary complex, dislodged from the plateau of Kagoshima (p. 25), and lastly, the fourth is a patch of ancient lava cap that covers the tliird (p. 26). In short, the writer is not in full accord with the view ex- pressed to the effect that the series of ventholes or bocclie are located on a traverse tectonic line in the fundamental structure of tlie region. There are a number of ventholes both in the front and the rear sides of Sakura-jima, of which some are only of ephemeral character {see p. 74, Text-fig. 19). Ventholes y^^ ^^.-^p cOUUt 5 lava veuts OU the frouf (wCSt) ON THE "^ ^ ■' ^\est ^-^i^^^ ^^^j ^I^^ ^^^ marked with a small circle is a mere epigonic blow -hole located on the edge of settling ground (Text-fig. 27, and a small cross, x, in fissured terraces in back- ground of Pis. XIV-XV.) in consequence of the defect of mass which escaped in the form of fluent lava. No. 1 (Text-fig. 19) at the nortli of the roundheaded Hikino-hira is a comparatively insignificant lava orifice, 30 m. deep, which might have been very 140 ART. 3. — B. KOTÔ Pig. 27. — Blow-liole nbove No. 1 venthole in the western lava-field, marked ■nith a cross in PI. XV. (Photo by Mr. Yamaguchi.) active during the first pliase. It is Danbrée's diatreme hole from one end of which lava flowed downwards (PL XV. No. 1). No. 2 — the Yuno-hira vent — is the giant of orifice at an altitude of 300 m. and really entitled to the name of crater. It was formerly a hollow alongside the upper limit of a cultivated gentle slope below the western precipice of the north cone, as may be seen in various photographic views prior to the eruption. .Vlready for a score of years, weak jets of steam have been issuing there from fissures in the ground. This time it engulfed a considerable hollow from which the lava shield rose {see Text-fig. 28) like a ' SchoUen- dom "Mn Kilauea,'^ and fluent lava broke through and forced its way down the slope (Pis. XIV.-XV.) pouring fortli nearly the wliole 1) The type of the enii)tion at Yitno-hira is exactly like that of Santorin in 186G when a dome was there first formed, but the lava (hypersthene-andesitc) finally burst oxat the side as a flow. — ^F. Fonqué, ' Santorin et ses eruption,' Paris, 187!). 2) F. A. Perret, ' Some Kilauean Formations.' Amer. Jour. Sri, 1913, p. 154. Till': (iRE.vT luirniON of sakuiia-jima in 1914. 141 Fig. 28. — L.avti-sliield of the Yuno-liira vent, viewed southwest ward s from the western slope of the North Cone. . (Photo taken on Jan. 19th, 11 a.m., 1914.) mass of the western lava stream, which deployed in its downward coarse like the linger of a hand and finally plunged into the sea (PL XII. Fig. 3). Nos. 3 (?) and 5 {sse\\ 74, Text-fig. 19) are probably secondary spatter vents broken open from the undercurrent of lava stream under solidified shell. Whether No. 4 is a primary or secondary vent is not clear to the writer, the balance being in favor of the former, which is probably connected with No. 2 in its deep source. Generally speaking, the western vents are rather irregularly 142 AET. 3. — B. KOTO : scattered and the paroxysmal activity was concentrated in No. 2 — the Yuno-hira crater. yentholes rj^^^Q j-f^Qj^ castcm ventliolcs, in contrast to tlie ON THE ^^^ irregular distribution of the western, are arranged in straight line {sec Sketch-map, Text-fig. 10) irrespective of topography, from northwest by west to southeast by east. This distinct alignment of orifices marks the trend and location of fracture in the superstructure of the volcano, and apparently justifies the hypothesis on tlio existence of a tectonic line in that direction through the whole mountain. During the first phase, the sight of the glare at night observed by the writer was a splendid one. When one of the vents, either No. 3 or 4, made an outburst, the rest readily followed and sympathized with the forerunner together with minor vents, and the general awakening of fiery torclies ensued one after another, finally fusing together in the form of a carmine-red wair^ of hellish blaze, as if hot lava was running along and underneath the fracture line (Text- fig. 14 [c]). Comparing the position of orifices noted during two trips, one in January, 1914, the other in April of the same year, some changes had occurred (Text- fig. 19) ; No. 1 and those marked with small circles on the southern slope of Nabé-yama had been rather active during the first phase, though they later became dormant. The order of their birth according to the writer's opinion was as follows : The activity started with the lava-flow at No. 1 (Text-fig. 19), tlie Sen-yemon recess popularly so called, which soon became ex- hausted, and then shifted to No. 2, and the one near by (not in- dicated in Text-fig. 19) was a contemporary. Nos. 3 (250 m) and 4 then followed, and displaj^ed vulcanicity on a grand scale with 1) A continuous glare appeared before the outwelling of fluent lava. THE GliE.Vr ElîUrTION OF SAKUJîA-JIMA IN 1914. 143 Pig. 29. — Activity of No. 5 vent on the slope of Nabé-yama, as seen sonthwest'nards from the Ebino-ziika hill. (Photo taken on May I7th. 1914.) brotherly sympathy when tlie writer was there in April, 1!)15. On the second visit (in April, 1914), the writer saw, greatly to his surprise, a gaping fracture opened between Nos. 4 and 5 on the southern shoulder of the pumiceous Nabé-yama homate (PL XI. Figs. 1 and 2), and two tongues of lava ran down northwards. No. 5 was a pecuhar epigonic lava venthole (Text-figs. 19 and 24), habitually standing in apathy to the others and remaining silent for long hours, spasmodically awaking with extraordinary violence and sharp detonations. During the writer's second trip, this bocca was active once a day, and was the most treacherous and dangerous one to approach (Text-fig. 29). The writer is unable to offer any satisfactory ox- Mechanism OF THE SaKDRA-JIMA Ij- n t • f»n j !• Eruption planatiou OU tlio mecliauism 01 the recent volcanic 144 ART. 3. — B. KOTÔ : activity (p. 137) ; therefore lie betakes himself to the realm of imagination. While in the field he was frequently reminded of the hypothesis of satellitic injection with which Daly'^ explains the in- different behavior of Kilauea and Loa. In Sakura-jima it was, however, bilateral satellitic injection (Text-fig. 30). For the duration of a full twelve months before the event at Sakura-jima, subterranean agitations, expressed in the earthquakes and the unusual attitude of the volcano Kirishima, were felt in southern Kyi'ishii, to release the sti-ain from an unstable isostatic equilibrium of the earth's supercrust. Lastly, a point of release was discovered in the old crater conduit of the south cone of Sakura-jima. Pig. SO. — Bilaternl iujection-chamber in Siikura-jima. (Horizontal and vertical relief in natural scale). Wolff {loc. cit.) says the local lava reservoir lies about 1,000 ?7i. from vent (Text-fig. 30), and on this assumption the lava basin of Sakura-jima will be reached at 700 m. below sea level. Driven np by hydrostatic pressure in conjunction with the liberation of gases which would have increased internal pressure and con- sequently adiabatic heat, besides that of new chemical recombina- tion, the lava ascended through channels up to half the height of the volcano by stoping, melting and gas-fluxing. Being unable to 1) • The Katnre of Volcanic Action.' Proc. Am. Aaid. of Arts and Sciences, Vol. 47, Jiüy, 1911. His diagram is inserted in Wolff, 'Der Vulkanismus,' Bd. I. S. 336. In Paly's own later -work, ' Igneous Eoc-ks and Their Origin,' it is not reproduced. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 145 overcome tlie burden of the higher part of tlic crater bottom, the magma sought easy and sliort ways of escape to the surface, where the mountain flanks i"emained yet uncovered with a lava coat of mail. ïlio naked slopes on both sides, the only weak portions of the whole mountain side, gave opportunity for the sites of surfiice activity by means of two series of ventholes, one on the front, the other on the rear. See Text-fig. 30. The ventholes were not in direct communication with the local main reservoir, but only indirectly, where lavas were injected sideways^ ^ through narrow and long passages from the main conduit of the top- crater. It is the satellitic injection chamber, and there were two of them which fed one or a number of vents overlying them, and in turn were fed from the main channel. The vents were mere openings of secondary injection chambers through a thin skin of rock-crust which separated them. That the sateUitic chambers were supported with magma fi-om the main channel is evident from the fact, that at the beginning of the activity threads of steam clouds ascended from the top-crater (Text-fig. 14 a [1]), which signahzed the immediate explosion on the western flank [2] and ten minutes later on the eastern (Text- fig. 14 6 [3]). While the lava made its way through melting up the crater conduit, partial gravitative differentiation took place, where- by the lower column of lava changed in its chemical nature to a heavy and basic one, and this lower column reënforced tlie sateUitic chamber of Nabé-yama (Text-fig. 30). The fluent lava of the east therefore consolidated to an olivine-bearing heavy and basic rock, while that of the west solidified usually without the above-mentioned mineral component. 1) It is the subnormal eruption in the sense in which Perret understands such a motle of subterrane lateral flow of lava. 146 AET. 3. — B. KOTO: ix. The Spatter Eruptions and the Formation of Domes.— The writer will say a few words on what he calls the spatter eruptions, a sort of Strombolian type, at the main vent of Yuno-hira on the western slope (Text-fig. 19, No. 2)— a scene described by Azunia^^ on January 22nd.^^ It was the actual scene on the 11th day after the first eruption. The crater pit was wide open and choked up with black vesiculatcd crust. A rent suddenly opened at one spot, and no sooner did the red tongue appear therefrom than a tremendous cannonading took place. Then a black curdy smoke (dust cloud) ascended, throwing up black lustrous fragments of fl.oating crust to a considerable height, which mostly fell again to the crater-bottom. The act was finished by emission of white (salmiac) fumes. After o or 4 minutes the same manifestation was repeated at another spot within the same crater. These spatter eruptions explain all the phenomena, seen from a distance, of the volcanic display since the 13th of January, and the writer believes this is the jKittcrn of activity of the volcanoes of andesitic magma. The above-mentioned Yuno-hira vent^^ also presents some remarkable features of uncommon interest. The venthole is of large dimensions, having really a claim to be named a crater. It is the main orifice which supplied nearly all the lavas of the 1) K. Azuma, 'The Great Eruptions at Salau-a-jima in 191i,' Kagosliima, 1914, p. 92. (Japanese) 2) See p. 87. 3) The above is Azuma's account of an adventurous attempt in peeping into the interior of the new active crater-pit No. 2 in Geologic Map and also Sketch-map, Fig. 19. This is a valuable contribution to the present paper ; for, what he has done is exactly that which the writer wished to have done. Azuma was the first who m.ide such an undertaking, which is often attended with risk. One investigator was mortally wounded by falling hot blocks at the cracking lava-front, and two others by unexpected explosions, one of them being the well-known vulcanologist, F. A. Perret. We have a ' Eeport ' from the pen of the American vulcanologist. • Eapporto preliminare sulla grande eruption del vulcano Sakurashima.' Zi'dschr, f. Vulkanologie, Bd. I. p. 133 d. seq. THE GREAT ERUPTION OF SAKURA-JEVIA IN 1914. 147 western field occupying an area of 7.202 sq. km. with volume of 0.2881 cub. kw. (p. 107). Sec Sketch-map, Text-fig. 19. A.t the western slope of the North Cone between the knob of Hikino-hira and the parasitic cone of Furuhata, there was a cul- tivated talus slope at the upper limit, at about an altitude of 300 VI., and there are a few ancient vents, one of which being the Yuno-hira, as may be seen on Map and in all the pictures of the western side. The Hikino-hira and Furuhata lavas are scars left from old wounds and the spots mark the weak portions of the old volcanic body. The slope between them is entirely changed in relief, a large cauldron having been formed with a margin of concentric terraces, each separated by gaping and fuming lines of fissures,^^ as can be distinctly seen in the background of the pictures in Pis. XIV.-XV. Then from the engulfed and collapsed bottom rose a gigantic * live ' dome, not unlike those of Tarumai of Japan and Bogosloß" of the Aleutian Islands. An excellent view of the smooth, bulging dome of lava from the north side in Text-fig. 28, taken 7 days after its birth, can tell the story bettor than words can depict. On seeing the successive slipping on the background of the figures in Pis. XIV.-XV. it soon suggested to the mind, that the solid ground was stoped and assimilated by ascending lava underneath, and the syntectic magma so formed contributed to the substance of the outwehed cupola. '^ 1) In a verbal commimicalion by Prof. F. Ômori, wlio visited the region in April, 1915, fifteen months after the spasm, he said that he had actually seen within one of the gaping fissnres here cited, a still fresh red-hot lava-ton// cf about 15 feet below. His important ob- servation gives weighty corroborative e%àdence in siipport of the writer's assertion that the lava in the local reservoir either stoped or assimilated the wall upwards, leaving only a thin margin of old surficial crust. Or, it may be assumed that the ramifying lava tongiies simply ran into open fissiires from a local reservoir by hydrostatic i^ressure. 2) It is not unlike Daly's " foundering " eruption. He says this type of eruption is con- ditioned by stoping and assimilation of bathoUthic magma, but in the present case we have before us the melting of the thin roof of a satellitic chamber. 148 AET. 3. — B. KOTO : The ephemeral dome was then ruptured on its southwest side through the expansion of lava and the pressure of the enclosed gases, leaving behind the hoof- shaped carcass of the dome, as may be seen in Plate XIV. on the left and the train of lava-flows from the ruptured side on the right in Plate XV. It was in the interior where Azuma observed the spattering eruption already referred to (p. 146). After the lapse of a year the feature of the Yuno-hira vent had entirely changed, the ephemeral dome had disappeared, as may be seen in the foreground of the picture in PL XII. Fi^. 3. In April, 1914, on his second trip the writer had a good opportunity of observing the process in detail from an isolated hill- top of Yebino-zuka near Kurokami on the east coast, when the vent No. 5 made a forcible outburst without warning but with tremendous detonation (Text-fig. 29), throwing up dark curdy smoke and grains of juvenile lava- crust. When the act was ac- complished, the lava gushed out down the slope with chnking as if a viscous fluid was pumped out with one stroke, often from a special lateral opening a Httle below the main vent, as may be seen in some of the Nabé-yama vents. The process seems to be as follows : The pent-up gases within the vent break open the plug or solidified skin of lava at the upper end of the conduit, resulting in the release of pressure in the column of lava. The occluded gases and steam then ex- pand and cool the magma wliich in turn becomes viscous. The liberation of gaseous bodies in the magma in consequence of relief of pressure tends to let gases free from the viscous body, causing thereby an increase of the volume of foaming, puffing and spatter- ing lava, and at the same time forcing up the upper lava column. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 149 This is the cause of the outwelling of it from the orifice (Text-fig. 29). It is a primarf/ lava-flow. X. Sublimation-products.— The activity manifested by nnmerons volcanoes in Japan is nsnally of the explosive type, and we rarely have occasion to collect the manifold products of subhmation as in the case of the recent eruptions of Sakura-jima, when fluent lava copiously effused from a series of vents. In saying so, it does not seem to agree with the fact that fumaroles and solfataras are by no means rare in this country ; but these mostly re- present the last feeble stage of vulcanicity, and the sublimates deposited are of oxidized and decomposition-products of low tem- peratures, corresponding either to the sulphureted- hydrogen fumaro- lic or to the solfataric activity (100^-200°C.), and consequently, meagre in variety. During his second visit, in April, 1914, the writer was able to approach the margin and also the solidified but stiU very hot lava-flows at many points. The clefts in the ancient lava near the Yuno-hira venthole were seen richly incrusted, exclusively with needles of bright yeUow sulphur — the direct crateral sublimate fi'om the lava reservoir, wliile at the edge of the lava stream, away from the vent, the dominant deposits were variegated lava- SuBu^TEs sublimates incrusting the ragged holes of ' block ' lava. The latter are of special interest to us, as we could not yet approach the lava mass within the vent itself. The incrustation consists of {a) a loose aggregate of white salmiac crystals, super- posed by {h) the yeUow crystalline layer of another mineral on irregular prominences, and lastly, by (c) the deep orange-red crystal aggregate on tips, as in the first snow on mountain-tops in early autumn. Unfortunately, a fine collection made on the spot has now 150 ART. 3. — B. kot5; become useless for microscopic and chemical studies, as most of the sublimates are deliquescent, and intermixed with intricate aggi'egates. (a) The main mass of the above incrustation is composed of salmiac.the salmiac, NHiCl, in the form of rhombic dodecahedrons, PBINCIPAIj Sublimate gometimcs flattened on the face (110).'^ The dodecahe- drons were picked out, dissolved in water to remove solid im- purities, and then evaporated to a white mass. This was analysed by Mr. Y. Yoshimatsu and Assist. Professor Y. Shibata of our University, giving 9 9. 6 OX of NH^Ol, besides a slight trace of Fe. The writer dissolved the same dodecahedrons in water and allowed the solution to evaporate on an object glass. Under the microscope the preparation proved to be skeletal crystals of salmiac, besides a number of characteristic rhomboid tablets of gypsum crystals. The gypsum is a never-failing compound in volcanic exhalation. (b) The yellow or orange -colored aggregate may be hexagonal molysite. Feds', or rhombic erythrosiderite, IKGl-FeCk'HiO. (c) The ruby-red tip-aggregate may possibly be octahedral kremersite, KGl-NE^Gl-FeCk + Sz 0. Y. Shibata made qualitative tests on a specimen sent to the wi'iter from Kagoshima. A mass of sublimates, found as incrusta- tions on a slaggy lava, is a white crystalline aggregate of gypsum, as can be readily recognized under the microscope, with a yellow coating. The other compounds were found to be the following : i. The solution extracted in water contains a small quantity of NHiCl, besides SOi and a trace of CI. ii. Treated with hot aqua regia, the solution shows the pre- sence of Ga, Mg, Fe (trace) and ICSO^. 1) M. MisAWA, Jow. Geol. Soc. Tokyo, p. 96, 1916. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 151 iii. Tlie aqueous solution after fusion with an alkaline car- bonate shows the presence of AlO,i and SiO.,, with a trace of Fe and Mn. The sublimates are composed of, enumerated in the order of quantity, firstly, GaS0,+2H.,0, then NH.CI, 3IgS0, and lastly, free SO,, The so-called rock salt with a saline taste, which tliinly coats the block lava, was dissolved in water and evaporated. Abundant crystals of gypsum in flat rhomboid tablets and dirty cubes of (?) rock salt constitute the dry preparation. M. Kanai found in a qualitative test made in fresh sublima- tion-products the following minerals with relative quantities in the order mentioned below : Salmiac, NHJJl ; kalinite, KÄl{SO^.{\2H.O ; molysite, FeCk \ chloraluminite, AlCliQH.j,0 ; rock salt, NaGl ; iron- alum, FeSO,-Al,{SO,),-24:H20. Summary To Summarize what is stated on these scanty data, the most dominant elements of the products of volcanic exhalation are undoubtedly salmiac and p r i m a r y gypsum,'^ besides sulphur. The cauliflower- shaped so-called steam column ascending from the vents and also the white fumes emitted by fluent lava streams can be reasonably attributed to the presence of abundant salmiac fumes (p. 07, footnote). 1) Here it is meant that the gypsum is a direct product of exhalation. 152 ART. 3. — B. KOTÔ : Part III. Petrography.^^ § I. The Lavas of the Pre-historic Cones. A. The Lavas of K-ita-daké. à) Topographic Features.— As it has been stated already (p. 35), Kita-daké or the ' north cone,' which carries the apical crater of Mihachi, 1,037 ??2., constitutes alone the main structure of Sakura- jima, making up the highest cone and occupying nearly the whole base of the island. See Geologic Map. At later times several lateral flows and subaërial talus, together with pluvial fans, only slightly contributed to the peripheral expansion which is at present about 4 km. in basal diameters. That the Kita-dakè lava is the principal building material can be confidently affirmed by tracing the occurrences of the same rock nearly through all parts of the island, although later flows hide and obscure exposures, especially on the eastern slope. •NoETHEEN Northern Slope. — A complete exposure (Geologic Map, n) from top to shore can be traced on the northern slope, terminating at the point of Wari-ishi-Saki. It is the Uno-hami^^ ßoiü. There is a lava-shelf or terrace at an altitude of from 200 to 300 m. on the slope, formed by stemming of subsequent flows of magma. The adjacent eastern area is now under the lava sheet of 17 79-' 80, while the western is overlaid by the lava apron {nk) of Kabano.'^ On the eastern skirt of the latter there are two narrow 1) The characterization of rocks under the present heading is merely of preUminary qualitative nature, detailed quantitative study being out of question at the i^resent moment. 2) ^/±. 3) See posteu, p. 158. THK GREAT ERUPTION OF SAKURA-3IMA IN 1914. 153 and dry channels through wliich torrents carrying pumice and ash once rushed down to the coastal villages. It is known as Yama-sliiwo or mountain-tide of the Matsura gulch (Matsura-gawara on Geologic Map) — the scars left fi'om excavation in 1779 (p. 40). The ravines lie petrographically in a weak line, being the margin of the Kabano lava ; and one finds in the bottom of the channel good exposures of slag-agglomerate and cap-sheet of the Kabano lava (p. 158). There is still another double gulch of minor size, lying to the northeast of the former, known as the Futamata gulch {see Map). Western Wcstem Slope. — The lava exposed on this side SLOPE •* •*■ appears firstly in the form of a parasitic cone — the volcanello of Furuhata^^ (Geologic Map, nf), and secondly, in a patch at Atago- yama (n), and thirdly, in a ravine of Ogawara (n). The space between the first and the second is overflooded by the recent lava, while that of the second and the third by the Ohira lava (oh) of 1749. In passing, it may be remarked that the pointed cliffs on the upper reach, called Jiasami or the ' scissors,' slid down in avalanche in the deep gulch of Hase in the early morning of January 12th, the result of constant trem^blings of the ground, and the slips were mistaken for the eruption, although the actual eruption began only about 10 a.m. of that day. On the cliff that bounds the deep gulch of the above-mentioned Hase valley (Mizu-otoslii in Geologic Map), strong agglomerate beds^^ 1) "Sr tffl (H ^ SM) The volcanello of Furuhata is an ancient adventitious tholoide of the North Cone, to which it is genetically related in the nature of the magma. Once a Avide open crater of the rheumatitic lava is still to be seen in the form of a fragment of wall, named Yunohira (Text-fig. 31, y), and a new low collapsed lava-dome (f) sits upon its western flank, which received the name of Furuhata. On the top (Sambongachi), there were three holes, constantly emitting lukewarm gases, only one of them still surviving after the recent paroxysm. 2) The clifE-profile shows the structure, beginning with a) the pumiceous éjecta and pumiceous lava, followed above by j8) the block bed with ashes, the whole being capped with 7) the lava-sheet. The total thickness amounts to 300 in., and the entire complex is built of the characteristic light-colored Kita-daké rock. 154 ART. 3. — B. KOTO : are, according to Uyeda and Yamagüchi, found exposed from under the lava-sheet (100 m. thick) of the North Cone. From this im- portant find it is evident, that the first phase of the activity of the Sakura-jima volcano was explosive, followed later by the outpourings of the rheumatitic material which built up the apparent tholoide (though really konide) of the present North Cone. The above-mentioned Atago-yama of dubious origin is a geologic island left untouched by the recent lava-flows (Geologic Map). The patch (n) of the Ogawara gulch is the inlier or fenster exposed through erosion from under the lava of the south cone. The recent éjecta of rock of this type (the north cone type) were found largely intermixed with the juvenile lavas on the western slope. The volcanic plug (?) of Ilikino-hira (Hyoku-oka). — A hornito- like boss elevates itself on tlie western slope to a considerable height (553 m.), though the northeast edge blends upwards with èà min-i^Hr i Kg-' WÊÊk Hi ^^^"""■^■'J El ' m w. . X ■«.n^^^^^^^l »■•..--, . <>-flt|[j||^ 1 Y V ^^^^^pHppPfS^«^^^ li^. Fig. 31. — The volcanello Furnliata on the western slope. (Photo by Mr. Yixmagnchi.) Yv — Yuno-hira vent. y — Yuno-hira somma. F — Collai^sed dome of Furuhata. H — Hakama-goshi. Ktj — Kanzé islet. k— Kagoshima. THE GEEAT ERUPTION OF SAKURA-JIMA IN 1914. 155 the north cone. All the I'ecent writers on Sakura-jima take it for a parasitic cone, simply from its outer form. Tt is built up of massive rock, which is of the saine land as that of the north cone. Until this monadnock-like mass is convincingly proved to bo a real parasite, the writer will relegate Hikino-hira to a part of an ancient lava-flow, which was greatly disrupted by explosions of old date in the neighborhood, so as to give the outward look of a steep dome or tholoide. Southern Side. — A small elevation, 10 o m. high, near the defunct strait of Séto, raised its head on shore fi'om below the ancient lava of the south cone or Moyé-daké. The writer passed over the edge of it and made a collection of the rock in January, 1913. This Akashi- Gongen hill and the whole adjacent tract are over- flooded by the recent lava and no trace of this "103 m. hill " can be seen now on the surface, except a slight elevation in the middle of it, indicating by the petrified cataract of recent lava the site of this buried hill. About the origin of the hill, we find no clue as to whether it is a margin of the North Cone or one of its parasites. (Geologic Map, PL VIII. Figs. 2-3, PI. IX. Figs. 2-3.) Eastern Side. — A knob of Gongén-yama (340 m.) on the eastern slope and the Sono-yama (Maru-tsuka or the ' round mound,' 79 m.) on the northeastern shore are the geologic islands of the Kita-daké lava or its parasitic cones left uncovered by later flows. b) Pétrographie Characters.— The Kita-daké lava (PI. XVI. Figs. 5 and 6) on the north side, a typical hypersthene-andesiteP is the oldest and the most widely distributed rock of the island. It is a light-colored, salic type, which rather resembles the Diluvial 1) As it contains sanidine in the forms of plienocrysts and the rind aronnd plagioclase, it may in future investigations turn out to be a hypersthene-latite or pyroxene-canLaUte. 156 ART. 3. — B. KOTÔ : and Tertiary efïusives (PL XVI. Fig. 1) of southern Kyûshû than the hlack historic volcanics of the island. The slaggy hollows in the rock are often secondarily filled with calcite and primary tridymite aggregate. It is a sHghtly slaggy, dopatic rock of light color, in various shades of gray, wiiich entirely depend upon different factors in the groundmass. The (a) brownish or purplish tinge is due to hematitization through fumarole action upon augite- needles and prismoids ; the {h) ash-gray tinge, which is very common, is caused by slight admixture of the microlites of augite and the crystals of magnetite in the plexus of feldspar microlites ; the (c) wet-grey comes from flocky aggregation of augite-anhedrons and magnetite dust ; and lastly, tlio {d) lightest shade and dull lustre is attributable to the predominancy of feldspar microlites and residual colorless glass. Phenocrysts of distinctly pleochroic pyroxenes, which are all of brownish green, are scarce and only represented by slender pleochroic hypersthene {h), but those of (/) labra- dorite (1-1.5 mm.) are fairly abundant. Simple anhedral feldspar, and idiomorphic poly synthetic - lamellar plagioclases as well, are usually zonal- structured. Glomeropor- phyritic^^ {g) secretionary patches of micronorite (Text-fig. 32) play a part of phenOCryst, being COmpOsd Pig. SS—Micronontic segregation. 1) Flocked together probably during the crawling of lava- streams. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 157 of an ophitic aggregate of plagioclase and pleochroic pyroxene ; the brownish hornblende, which was only once seen in the Saknra-jima shdes, occurs marginally'^ intergrown with the pyroxene. Apatite was observed once or twice in slides. In the lavas of the southern slope augite is represented in moderate quantity. The texture of the groundmass (PI. XVI. Figs. 5 and 6) is pilotaxitic and hyalopilitic (if colorless glass is abundant) and fine, and appears dirty gray under weak powers (Fig. 5). As already stated, the color of the rock is solely influenced by different be- haviors and a relative quantity of microlites of feldspar and augites besides magnetite and glassy base in the groundmass. One exception to the general rule is the lava on the northern shore at the Point Wari-ishi-Zakl near Kômen. It is stuffed with light- colored, crystalline nodules of secretionary origin ; but the lava itself is blackish, being built up of broivn, sometimes colorless, glass in which only resorbed and dasted augite prismoids are seen swimming. No feldspar was detected among microlites, owing perhaps to rapid cooling under water, which did not allow the crystallization of feldspars. The Atago-yama lava is gray with brownish flecks. Micro- scopically, it is similar to that of Wari-ishi-Zaki, having networks of chilhng cracks, appearing as if it were sohdified lava under shore water. A noticeable feature is the colorless devitrified fibrous glassy base with corroded and fibrillated augite microlite, which is partially oxidized into hematite dust. Brownish flecks are solely due to this dust, which is also strewn through the oxidized margin of slaggy holes. 1) Bowen says in his experiment ou the gravitative difiEerentiation of melts, that olivine and pyroxene during their settling (olivine-nodiile) cause crystallization of amphibole upon their cooling surface. Amer. Jour. ScL, Vol. XXXIX. 1915, p. 175. 158 AET. 3. — B. KOTO : The lavas of Fariihata-SamhongacJd (PI. XVI. Fig. 7) and Hikino-hira. — They are all as the above with few hypersthones and abundant plagioclases. The groiindmass is also light-gray under weak powers. Higher powers resolve its hyalopilitic texture approaching to that of pilotaxitic. Rectangular and lath- shaped twinned plagioclases make up the main bulk of groundmass, intermixed with a subordinate quantity of fine needles of augite, besides grains of magnetite which are found mostly attached to the augite. B. The Kabano Lava. (North Side.) PI. XVI. Fig. 8 ; PI. XVII. Fig. 1. a) Mode of Occurrence.— An old lava descended the north- western slope to the 200-meter contour-line, from the interior of the (?) apical crater (terminal lava-flow). It hangs on the slope or piano in digital lobes and forms a characteristic shelf at its end bv stemming and coolino; of the viscous lava stream in its Fig. 33.— North Cone (Kita-dal;,') -xiewecl from ncwr Sbira-hania. Cf. Text-fig. 37 (p. 173). 1 — ^Kabano laAa-flow. 2 — Erosion-gulcli of Matstira-gawara. 3 and \ — East and west ventlioles of the An-ei eruption. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 159 downward course. Tlio northwestern rim of the Kita-daké is built up of this black glassy lava. Geologic Map, and Text-figs. 33 (1) and 37 (1). The relative age of this rather weak lava cannot be stated with certainty. It solely rests upon the Kita-daké lava, and is not in anyway related to other flows, so that it is still an open question whether its eruption happened soon after the building-up of Kita-daké, or later in historic periods. The writer puts it on record here merely fi'om a geographic point. As stated elsewhere (p. 153), the eruption began with the ejection of ash and pumice, followed by and covered with the later flow. Exposures are seen in the gulches of Matsura-gawara. Text-fig. 33 (2). From a pétrographie point the lava belongs rather to historic types. In contrast to the light- colored rock of the north cone, it is black and vitreous, dotted white with crystals of feldspar (from 1 J mm. downwards). The rock is hyalopilitic pyroxene (hyp>aug) andésite. h) Pétrographie Characters.— Under the microscope, this dohyal- ine and dopatic rock is seen to be largely made up of colored glass (often colorless), with shades of purplish-brown, dark- brown and red -brown, variously kneaded and streaked after a pattern of damask (Fig. 1). Abundant augite in the form of fibrous and also minute stiff needles in tufts and axiolites (Fig. 8), and relatively few skeletal plagioclase are present in the brown glassy base. Phenocrysts of zonal- structured plagioclase are abundant ; but pyroxenes are few, of which the prismatic hypersthene {h in Fig. 8) predominates over anhedral augite {h in Fig. 1). IMagnetite is present in small quantity. The lava is a black vitrohypevsthene- andésite with augite -microlite or brown glass in the groundmass. 160 ART. 3.— B. KOTÔ : C. The Minami-daké and Naka-daké Lavas. (South and East Sides.) PL XVII. Figs. 2, 3, 4, 5. a) Mode of Occurrence.'^— Although Minami-daké (the south cone) and Naka-daké (the middle cone) are really parasitic to the northern main, the first can match the northern in point of magnitude, and rides on its southern flank. At the first phase of the building-up of this brother giant to a lofty elevation of 1,070 -??!., the southern shoulder of the old Sakura-jima volcano must have been already bodily blown ofi' and destroyed when éjecta of preexisting material and of juvenile substance were piled up around the vent of the south cone. Lavas alone without the contribution of breccia'^-* and tuffs cannot form a priori such an overtowering konide. The deep gulch on the west side seems to be the overlapping edge of both the north and south cones, where one expects to find their mutual relations and internal structure. As may be seen on Geologic Map, the black Minami-daké lava makes up the entire apical region, while its skirt borders the southern shore on the point of Kannon-zaki between Yuno and Furusato, and further east at Yuno-hama, being interrupted midway by the lava stream of 1779. Besides, the southwestern slope is severally overlaid by the Biyôbu-hira^^ lava of 1475-'76, and the Ohira^^ lava of 1749, and between the two is exposed the infier or wmdow of the old Kita-daké lava at the gulch of Ogawara.^ These three periclinal lavas of flank eruptions complicate the topography and distribution of lavas, both in time and nature. 1) See ante, p. 35. 2) Strong clasmatic deposits are exposed along the southern and eastern slopes between the 200- and 400- meter contour lines. 3) # JSi ^ See postea, p. 166. ^) :k ^ (spolœn U'hira) The Jam means a flat. 5) See ante, p. 154. :>*C fPl J^ THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 161 Naka-daké Tlio eastern sleeve of the Minami-daké lava-field is sheeted with a flow from the central eruption of Naka-daké, which covers the eastern slope and marks also the coastal strip of Nagasaki-hana or headland. The already -mentioned excentric Nabé-yama and the minor Ebino-tsuka ringwall of 1471 sit upon this base, as may be seen on Geologic Map. It should be remarked, that at the xA'esent state of the writer's knowledge it is doubtful, whether it is practicable to dif- ferentiate the Minami-daké and Naka-daké lavas or not. {See Text-fig. 9, M.) The top-region is so thickly coated with blocks, lapilli and ashes of historic and prehistoric eruptions that ex- posures of underlying rocks cannot be seen, except at the deep inaccessible radial gulches, which are few in number in this still undissected volcano. On Geologic Map the hypothetic area of the Naka-daké lava is brought to cartographic expression in order to suggest a problem to be worked out in future researches. In passing, it is to be remarked that there may be another point of efiusion of magma near Nabé-yama, adjacent to its northwest skirt at a knob of 340 m. It is Gongén-yama, and it may be a part of the Kita-daké or a volcanello belonging to it. b) Pétrographie Characters.— i. The specimens (PI. XVII. Fig. 3) collected on the southeast slope between 300 and SOO m. are black, dopatic and compact pyroxene- andésite. Phenocrysts (1.0-1.5 mm.) of andesine plagioclase are abundant, but pyroxenes are as usual scarce, of which prismatic hypersthene predominates a little over anhedral augite. Phenocrystic pyroxenes are by no means easy to distinguish and show abnormal features. Hypersthene is mostly highly-birefringent, and augite is mostly pleochroic. Both have exactly the same brownish green color. Prismatic section with 162 ART. 3. — B. KOTÜ : eight-sided outline affords, however, a good diagnostic mark to hypersthene, wliile the tabular (100) with twdnned basal section that of augite. Parallel intergrowth with peripheral augite was sometimes observed, but the reverse was only once noticed (S. E. slope). The statement on the relative quantity of the two ]3yroxenes should therefore not be taken in a strict sense, especially the phenocrysts are not found in large quantity, as in the lavas under question. Simple subhedral tabular (010) feldspars as well as polysynthetic-lamellar, long- rectangular plagioclase are zonal- structured. The hyalopihtic groundmass is fine and built up mainly of minute augite needles and lesser and larger laths of feldspar in brownish glass base with magnetite crystals. Specimens fi'om the neighborhood present textural varieties ; some l^eing brownish and slaggy, the other being fine in texture. But the mountain slope is thickly covered with all sorts of ejected blocks of several erup- tions, so as to make a proper selection difficult in collecting specimens. ii. The lavas in the Kannon-zaki area (PI. XVII. Fig. 2) on the southern shore are likewise black and rather vitreous owing to the abundancy of glass in which the microlites are mostly re- presented by augite, besides a few larger feldspar-laths. Olivine was once noticed, and micronorite patches were also seen. iii. The Naka-daké lavas, which cover the («) top and eastern slope and the (ß) coastal strip of the Nagasaki headland, arc rather slaggy and colored brown in vesicular spaces through oxidation of the pyroxenic components. The « typo is a jntch -black vitroandesite having a groundmass of abundant brown glass with augite needles only. It is a porphyritic obsidian (PL XAMl. Fig. 4). The ß type (PI. XAMI. Fig. 5) is hyalopilitic and diiU-lMck THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 163 dae to the predominancy of fine feldspathic microlites over those of augite in light-brown glassy base. Both types are found inter- mixed, a fact which seems to justify us^ in putting both together in one group. § II. The Lavas of the Historic Period. A. The Lavas of the Bummei Eruption/^ 1471-1476, Including the Nabé-yama and Ebino-tsuka Adventives. «) The Urano-maye Lava-field of 1471. (I^iist Side.) PL XVI IT. Figs. 2 and 3. a) General.— Taken in a broad sense, the grayish and salic, porous and light lava of Kita-daké, which, constitutes the general foundation of the insular volcano of Sakura-jima, difiers fi^om all the later lavas, irrespective of historic or pre-historic, which are dark and basic, slaggy and heavy. The femic jwe-liistorlc lavas already briefly outlined as the Kabano (?), Minarai-daké and Naka-daké lavas welled out in streams, or built up mighty cones during the mature age of the volcano to a height of more than 1,000 m. Although dark in color, the three lavas are r/u//-black, owing to the predominance of feldspathic microlites in the groundmass, oi' else of light- brownish color of glassy base. Among the pyroxenic phenocrysts, Ivypersthene usually predominates over augite^ indicative of a ratlter salic nature of the magma, representing a transitional stage to more femic historic lavas to he considered immediately below. 1) See (tiitc, p. 42. 164 ART. 3. — B. KOTÔ : All the historic lavas, which are likewise dark in color, are fresh pitch-hleLck, due to the presence of dark or violet-brown base, and they are the product of flank eruptions, in contrast to the central, showing that Sakura-jima has already passed in pre-historic times the prime of its volcanic age. b) Mode of Occurrence.— It is stated^-* that ' on October 25th, 1471,^^ the mountain was in flames above the village of Kurokami (on the east coast) projecting stones and raining ash. Lava piled A up and formed a rocky hill called Omoyé-zaki, the great burning cape.' With this simple passage as a key, the writer has differ- entiated a lava field at the north of Kurokami from that of 1779, which encloses it on two sides, leaving a triangular uncovered patch of 2.4 sq. km. along the shore. See Geologic Map. The field lies 1.2 km. north of Kurokami, and this triangular lava terrane has 3 depressions at the apex landward, while the pro- jecting headland of Omoyé-zaki forms another corner. The latter is the visible terminus and the former the vents from which lava poured forth seaw^ards to the above-mentioned headland, and which people call the fire-holes or higona of Urano-mayé or Uran-mayé, the latter a solitary group of cottages in a cove. The ventholes of the Urano-mayé lava-field are the fresh- looking depressions within a terrane of lava clods, the largest being elliptical with axes of 200 m. by 100, and bounded with steep cliffs 50 m. deep. The bottom of the pit is covered with talus, and old ashes. The black lava-field is craggy, affording foothold only to stunted trees and meagre grass. Topographically the writer is able to recognize this Urano-mayé lava-field of 1471 1) Chiri-.mnh o Jfi 51 ML^- ^«rt XX. p. 12. ' Geographic Scraps of the Provinces Hyiiga, Ôsumi, and Satsuma,' KHgoshima, 1898. 2) Some say it occurred in 1475 ? THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 165 from the enclosing one of 1779 in its being of low altitude and fi-om the tradition of the people, especially the formation of Omoyé-zaki or the ' burning headland ' of whicli we have an authentic record. c) Pétrographie Characters.— The material of lava (PI. XVIII. Figs. 2-3) is white-spotted (1-4 mm.), dopatic, slightly slaggy j^'^tch- black rock which, when seen under the microscope, is exactly like the younger lavas of 1779, later to be described. It is a pyroxene- andcsite with the groundmass of dark brown glass in which rather coarse microlites (0.06 mm.) of dusted augite and skeletal plagio- clase swim in fluidal arrangement. The relative quantity of the two species of microhtes is variable, although that of augite always predominates over the other. Pyroxene phenocrysts are scanty and both members are variably represented. ß) The Nabé-yama and Yébino-tsuka Homates. (East Side.) a) The Nabé-yama Homate.— Contemporaneously with the activity and the outpouring of lava in 1471, Nabé-yama seems to have been also active. Mr. Ijichi in his oft-cited work : ' On the erup- tion of Sakura-jima in the An-ei era,' incidentally mentioned that the kettle (nabé)- shaped Nabé-yama is a relic of the eruption in 1471. He compared its position to the main Sakura-jima just as the parasitic knob of Hoyei-zan is related to Fuji-san. Particulars about its activity are not recorded in any work at the writer's disposal. (Geologic Map, hn.) This model-like ringwall hill is located on the southeast slope and rises on a base of l.?)hn. in diameter 328 ??2. high. The outer slope of regular ringwall is tolerably steep, but the deep inside is still more precipitous, and the circular wall is only 166 ART. 3. — B. KOTÙ: opened on the east, from which one can see the pumiceous bottom through the densely grown shrub. Nabé-yama is a unique, widely-opened kettle hill, entirely built up of stratiform loose lapilli and pumiceous mass. Therefore, it is in a strict sense a true homate. The loose covering of éjecta is in part no doubt of later origin. Hills like this both in form and substance are not unfrequently noticed in submarine volcanoes, the typical example being the new island near Iwo-jima on the south of the Bônin islands, which rose on January 23rd, 1914 and later disappeared. The reservoir of magma must have been near the surface, not lower than 300 m.; the juvenile material being all lapilli and pumice, formed and ejected by gaseous evolution in the magma. Solid lavas have so far not been discovered, and a comparison with those of the same eruptive period in other quarters of Sakura-jima is impossible. h) The Yebino-tsuka Knob.— A hiUock {see Geologic Map, h n) which lies to tlie east of Nabé-yama is an exact copy of the latter both in form and material, differing only in size. The writer is inclined to consider this pumiceous bocca to be of the same age as its neighbor. This little hill, though half enclosed by the recent lava flow, is so favorably and safely situated that all the visitors land- ing on Kurokami went up and saw the recent terrific display of vulcanism from there (Text-fig. 29, p. 143). B. The Byobu-hira Lava Field of 1476-'76. (Southwest Side.) PI. XVn. Figs. 6, 7, 8 ; PL XVIII. Fig. 1. a) Mode of Occurrence.— As the lava reservoir of the Bummei era seemed not yet to have been exhausted in 1471, activity revived THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 167 five years later, this time on the diametrically opposite southwest slope {see ant 2, p. 42). On September 15th, 1475, a violent ex- plosion occurred above Yuno at an altitude of 400 m. Another version says that the fluent lava issued forth on October 6th, 1476.^^ Lava welled out in streams, one branch spreading itself southeast- wards on the mountain slope in an apron, while the southwestern arm started from the vent,*^ No. 6 {t) of Geologic Map, ran d( -wii directly to the shore and pushed further into the sea, thereby creating the new point of Moyé-zaki (Text-fig. 4, cm.) or the ' burning headland.' The area overflooded by block lava is a little gi-eater than that of the earlier flow of Urano-mayé (p. 164). This Byôbu-hira field'*^ still leaves a series of six perfect vents or Ixx'cas (boche), which hitherto have l,>een unknown to the outside w^orld. No. 1. The uppermost pit'' (Text-fig. 34 [1]) is the typical and the largest one, located at a height of 400 m. The inner diameter is 30 m. and the circumference 200 m. with a depth of 40, the northwest rim being a little lower. As it is a blow hole, the back side is clear-cut in the old ash-gray lava of Kita-daké, a continuation of the Ogawara inlier (p. 154), and the fluent lava escaped only over the low^er northwestern rim, as in No. 1 venthole of the w^estern lava-field in the recent eruption (Text-fig. 27, p. 140). On the cliff" in the hollow (Text-flg. 34) a fresh slag- agglomerate 1) See page 42. 2) The vent No. 5 is named Takachiki (;^ ^]. 3) ;^ Mi ^ :^ II?f A view of the ByObu-hira lava tielil may be seen in PI. XI. Fig. 3. 4) The four pits from Nos. 1-4, counting from above, received the names Moyi-gowi (-^ ^ or the ' burning bocca '), Shitano-moyéi/oivi (the ' lower bocca '), llehiga-Jmljo {^ if *^ or the 'snake's hollow') and lastly, SurUxichino-kiibo (^^ #|v / -^ or the ' mortar hollow ') respectively. The road to the boccas leads up directly from Yuno, first entering into a block lava-field where there are tico wind holes (the vents Nos. 4-5 in (Teologic Mai>) in depressions. A cellar is especially built in each the above-mentioned depressions for the storage of egg-papers of silk worms, as the cool and constant temperatures of ' blowing caves ' are favorable in keeping back the untimely hatching of eggs. They seem to be terminal openings of lavantunnels with which the above-mentioned two lower secondary ventholes (Nos. 3-4) had probably direct communication. 168 ART. 3. — B. KOTü is seen to full advantage, overlaid by the sheet of the Kita-daké lava, that constitutes together with the underlying the main body of Sakura-jima. It is an excellent exposure to get a glimpse into the structure and rock- nature of old lava. No. 2. The next circular bocca (Text- fig. 34 [2]), 64 m. dis- stant from the first, is rather small with a Fig. 34. — Vents Nos. 1-2 of the Byôbii-hira lava-field of 1475-1476. Cf. Geologic Map. diameter of 130 m. and a depth of 40 m., but is a typical deep clear-cut and impenetrable venthole, this time of new black lava, densely forested with Cryptomeria japonica. One cannot find the spot unless one stands on the very edge of the pit. No. 3. The third pit in the series lies at about the same distance from No. 2, as the latter is separated from No. 1. It is a shallow basin (Text-fig. 35 [3]) filled with éjecta, which are also spread about over the neighborhood. No. 4. The last and the lowest pit (Text-fig. 35 [4]) lies a little apart from the preceding No. 3 pit. In its form and forested condition it closely resembles No. 2., the only difference being the smallness in size and the deep bottom (100???.). These four ventholes are arranged in a series on the slope in THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 169 Fig. 35.— Vents No. 3 :xud 1 in tlie same liiva-tiekl as in Text-fig. 34. the lava-field called Byôbu-hira or Shibano {see Geologic Map) in a south by southwestern direction. Though comparatively small in size, especially Nos. 2-4, these ventholes must have displayed tremendous phenomena in detonating and throwing up cloud dust, just as recent ones have done and are still doing on both flanks of Sakura-jima. We find no particulars of their activity on record. Tlie v/riter is not cei'tain whether all these represent the vents on the body of the mountain. Some may be secondary vents formed by the escape of accumulated gases within the fluent lava stream. A well-defined venthole lies between the 100- and 200-metre counter lines. It is the Takachiki vent (Geologic Map, No. 0, i^), and it probably poured forth the fluent lava which ran down to the sea forming Cape Moyé-zaki. b) Pétrographie Characters of Pyroxene-andesite.— The lavas are identical with those of Urano-mayé (p. 165), although they occupy 170 AKT. 3. — B. KOTO : areas on diametrically opposite shores, and in point of time both are only separated by an interval of five (?) years, dnring which no magmatic differentiation seems to have taken place in the lava reservoir, which is presumably common to the two lava-flows, as in the recent lavas. The common feature of both the Urano-mayé and the Byôbn- hira lavas is the development of varietal modifications in different parts of the lava streams. All the lavas are, however, dopatic or sempatic, dotted with abundant phenocrysts of plagioclase (1-3 mm.), while pyroxenes (hypersthene>augite) are only sparsely re- presented. The land lavas (PI. XVII. Figs. G-7) are pitch-black and vitreous, while those at the water's edge (PI. XVII. Fig. 8, PL XVIII. Fig. 1) are dull with shades of gray. The first is caused by the abundancy and chocolate -brown color of glass and also dusted augite-needles, while skeletal microlites are few. The sJiore lavas present the same feature as the land lavas in regard to phenocrysts ; but the groundmass is rich in feldspar -microlites, which lend to the rock its dullness and gray tinge. Moreover, the shore rock is cloddy, owing to quencliing fractures, and is stained with limonitic films. Under the microscope circular minute flecks are also seen, which are decolorized in its glassy base with the simultaneous separation of fibrous augite (enstatite-augite ?) and globulitic substance. These features are characteristic to shore Icwas irrespectiüc of age and chemical composition of the magma. C. The Öhira-yama Lava-flow of 1749 (Westeen Slope). PI. XVIII. Figs. 4 and 5. a) Mode of Occurrence.— As a precursor of the gi'eat eruption of the An-ei era (1779-1799), ' Mt. Ôhira-yama (spoken Uhira-yama) THE GREAT ERUPTION OF SAKURA-JTMA IN 1914. 171 above Nojiri erupted violently in the second year of Kwan-yen era (1749) ' after remaining dormant for four centuries.'^ On the southwestern slope just on the south of the mamelon-like knob (553 m.) of Hikino-hira (Hiyoku-oka),'^ the writer found a doubtful bocca at an elevation of 500 m., and another, but fresh and small one, at a spot 100 m. lower than the preceding. See PL XI. Fig. 3. This lower nameless shallow vent, which here may be con- veniently called the Hikino-taki (cataract) bocca, lies at the head of a lava stream which flowed down for a distance of 2.5 km. in a tongue- shaped strip without reaching the shore, and is partially overflooded in its northern skirt by the recent lava. See Geologic Map. b) Pétrographie Characters of Pyroxene-andesite (PL XVIII. Figs. 4-5). — The white-spotted (2-3 mm.) dopatic lava bears the stamp of historic type in its dark color ; but it is I'ather compact, break- ing in conchoidal masses with cloddy texture. Moreover, the groundmass of this black rock has a dull, lustreless gray shade, due to the large amount of rectangular and lath- shaped feldspar swimming in a light-brown glass. Microscopic flecks in the groundmass are caused by local globulitization and decoloration of brown glass. Pyroxene pheno- Fig. 36. — Micronoritic segregation in the (Jhira lava of 1749. 1) See page 43. 2) Hiyoku-oka ib ^ [ü^], an old name of Hikino-liira ^| y ^. 172 AKT. 3. — B. KOTÔ : crysts are as usual very scarce, but augite predominates over hypersthene. Excellent micronoritic segregations are typically developed in this lava, as in Text-fig. 36. D. The An-ei Lava of 1779 to 1781. (Southwest and Noetheast Sides.) PL XVIII. Figs. 6-8, PL XIX. Figs. 1-2. a) Mode of Occurrence. — After a lapse of thirty years from the slight eruption of Ohira-yama in 1749,^^ the most terrific paroxysm ever recorded of Sakura-jima happened in the An-ei era, especially from 1779 to 1781. This An-ei eruption has left a more dreadful impression than that of the Bummei era from 1471 to 1476.'^ Like the preceding activity, the An-ei eruption occurred also bilaterally in the same orientation, i.e., on the southwest on one side and on the northeast on the other, as may be seen on Geologic Map. The vents of the former period lie at a compara- tively low elevation of 400 m. in the Byobu-hira area, while those on the opposite side at Urano-mayé are probably hidden by later flows, the vents exposed to view being only at about 200 ??z. above sea-level. In contrast to the former, the vents of 1779 lie far up at an altitude of 700 m. on both sides, corresponding to the greater amount of energy exerted in the volcanic manifestation tha7i in that of the Bummei era. Although the eruptions were active on botli slopes, as has been already stated, the lava emission on the south (the large vent of An-ei- San, PL XI. Fig. 3) is comparatively insignificant. 1) See page 170. 2) See page 163. THE GREAT ERUPTION OF SAKURA-.TIMA IN 1914. 173 as compared with the vents of the northeast (the two^^ ventholes of the Moyédashi fi]'e holes, in Text-fig, 37 [3 and 4]) which not only sent down lava into the sea but also its submarine end rose to become the islet of Inoko-jima (a Rock off Osé-zaki in Text-figs. 1 1 and 12). A group of islands comprising Shin-jima, Iwô-jima, and Yébisu-jima represent, however, independent excentric submarine vents (Text-figs. 11, 13[y, i, s]), which were undoubtedly connected with the same lava macula of Sakura-jima, and sympathetically burst open with the activity of the main conduit. Fig. 37.— North Cone (Kita-daké, K) vie^ved from Kômen. Cf. Text-fig. 33, p. 158. I Photo by Mr. Yamagiichi.) To speak more in detail, the southern lava issued from a single deep and open oval bowl, called An-ei-San, on the apical region of the south cone, running down in a narrow tongue to the shore, and forming the point of Tatsu-zaki between Furusato and Yuno-hama — a stretch of only a kilometer. The fresh craggy stream of block lava can be easily recognized even from a dis- tance. See Geologic Map. The northeast area, the Komen"^ lava-field, is ten times larger than the preceding. The lava welled out from the east and 1) There are altogether eight ventholes, primary as well as secondary, in the Moyédashi lava-field, marked in Geologic Map. 174 AKT. 'ô. — B. KOTÛ : the west higona or fiery blast holes (Text-figs. 37 [3-4], 11), besides the third, which lies a little lower. The vemtholes lie at about 700 ??^ above sea-level. Three other bocche, the fourth, the fifth, and the sixth,^^ are located be- tween the 200- and 300-meter contours on the shelf formed by stemming of vis- cous lava-flows, and they are pro- bably secondary vents, burst open by local ac- cumulation of gases in a tunnel Fig. 38. — The sixth venthole in the northeast lava-field of 1779. See Text-figs. 10c, 11, :53, 37 and (reologic Map. within the cool- ing lava stream. The Komen lava spread out into two branches overwhelming the once large village of Mukômen {see Text-flg. 11), and enclosing the triangular area of the old Urano-mayé lava-field^^ of 1471. h) Pétrographie Characters of Hypersthene-andesite.— «) The lava of An-ei-San has the same black appearance as the other historic lavas, differing only slightly in lustre and in the sporadic oc- currence of olivine— n. component which is for the first time plainly observed in the lavas of Sakura-jima (PI. X\7n. Fig. (5). The lavas are all of hypersthene-andesite. 1) The last lies on the roadside, 'J'ext-fip. 3S. 2) See page 163. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 175 The rock at the water's edge is weakly vitreous, dark- gi'ay, somewhat slaggy and dopatie. As usual, the phenocrysts of andesine-plagioclase are abundant, whik^ those of pyroxenes com- paratively few, of which the prismatic hypersthene slightly pre- dominates over anhedral augite. The hyalopilitic groundmass (PI. XVIII. Fig. 7) is built up of dusted augite needles, mixed with a slightly smaller quantity of feldspar laths in a light-brownish glass. Microscopic flecks are seen, being caused, as usual in subaqueous lavas, by globulitization and decolorization of the base. The lighter shade of rock is caused by the minuteness of size but not the quantity of augite microlites in the groundmass. The lava on the slope is of a grayish black type with slaggy sempatic fabric. The phenocrysts are likewise plagioclase and pyroxenes, the latter being comparatively abundant as com- pared with the other historic lavas. The dominant pyroxene is hypersthene. Anhedral olivine again occurs sporadically. The glass base is colorless, but if the brown glass partakes of the groundmass in appreciable amount, the rock becomes vitreous and black. ß) The lavas of the Kômen area (the N.E. coast) are also divisible into those on land and shore. The former contains a comparatively large amount of hypersthene. The groundmass (PI. XVIII. Fig. 8) is full of augite microlites, intermixed with a little larger feldspar iu the chocolate-brown glass. The latter is characterized by minute texture (PL XIX. Fig. 1), of which there are two types, one variety Ijeing dark with a brown glassy base and augite microlite, the other being gray owing to very fine augite needles in the colorless glass. The common feature is the oc<;asional appearance of anhedral olivine and the decided pre- dominance of hypersthene over augite among phenocrysts. 176 AKT. 3. — B. KOTÔ: y) The rocks of the new islets oif the northeast shore (Text-fig. 11, p. 49) are ah-eady outhned elsewhere.^^ The sub- marine lavas, of which they are bnilt up, are gray or blackish gray, dull and pumiceous in color and texture, and not unhke an imperfectly coked brown coal. The groundmass is characterized by a frothy brown glass, in which fibrillated polysynthetic augite needles (clino-enstatite) and a few skeletal plagioclase laths are mixed with magnetite dust (PL XIX. Fig. 2). It is to be remarked that hitherto no olivine was detected, except in the lava from Iwô-jima. The characteristics of the An-ei lavas are the sporadic ap- pearance of anhedral olivine and the abundancy of pyroxene - phenocrysts, especially in the Kômen area, of which hypcrsthene abounds in preference to augite. These striking features distinguish the An-oi lavas from aU the ancient lavas of Sakura-jima. §111. General Pétrographie Characters of the Lavas of 1914. A. Massive Lavas. PI. XIX. Figs. 3-8, PI. XX. 1-4. a) Land Lavas in General.— The ' live ' lavas, stiU hot when collected, are exclusively the so-called block lavas or of the ' a a ' type, and petrographicaUy belong to hy per stliene- andésite, with olivine as an accessory. The writer can find scarcely any difierence between the recent lavas and those of 1779. They are more or less slaggy, sempatic to dopatic, and raediophyric. 1) See p. 54. THE GUEAÏ ERUPTION OF ÖAKüItA-JIMA IN 1914. 177 Phcnocrysts Jiro tabular lalmidorite crystals (2 7nm. or less), seriate and zonal structured with difterent extinction-directions, one variety being clear (PI. XIX. Fig. 3), the other being full of glass inclosures (Fig. 4) ; some are corroded and rounded in outlines, and often internally ruptured. Microphenocrysts are pyroxenes, of which the prismatic pleochroic hypersthene (PL XIX. Figs. 0, 7, PI. XX. Fig. 4), octagonal in l)asal section, usnally dominates over anhedral augite,^^ and both are often found in parallel intergrowth only in lai'ger individuals. Grayish apatite occurs as an accessory. The other microphenocrysts are the light greenish -yellow olivine and occasionally magnetite clumps, the olivine be- ing never idiomorphic in thc^se fluent lavas (PI. XIX. Fig. 7,0 ; PL XX. Fig. 3). The olivine'^ habi- tually forms a nucleus of glomerophyric mass, o,- the outer zone being a divergent mass of hypers- thene prismoids (Text-fig. 39, h), and the amoeboid iron ore constitutes a transitional zone. The whole appearance sug- gests the idea that the Fig. 39. — Cnxcked olivine core with a rim of hypersthene crystals in the land-lava of the western field. 1) The crystal is tabular on the pinacoid (100), and the random sections usiially met with are also the same pinacoidal plane with straight extinction, so that the estimation of the relative amount of both the straightly and obliquely extingiushing pyroxenes is only approximative. PI. XIX. Fig. 8; PI. XX. Figs. 1 an.l 3. 2) See posted, page 18.5. 178 ART. 3. — B. KOTÔ: intratellnric olivine was resolved^^ into iron ore and liypersthene (Text-fig. 40), and, chemically speak- ing, olivine is equal to iron oxide plus liypers- thene. At times pseu- dopodiform iron ore is simply left, forming the coi'e of the hypersthene- aggregate (Text-fig. 40). On another occasion, feld- spar laths enter into the pjg 40.— Hesorbecl olivine with pseudopodiform iron composition, building up °^^' ^""'"''^'^ ^^ hypersthene crystals in the ^ o ± marine lava of the eastern field. micronoritic patches (oli- vine-norite) and motex. Lastly, an isolated anhedron of olivine makes its appearance, and in tliis case it is not easy to discrimi- nate this mineral fi'om tabular hypersthene. The groundmass is always of minute fabric and hyalopihtic, being built up of prismoids or microlites of augite with a subordinate quantity of feldspar laths and magnetite grains in bi'ownish glassy base. See PI. XIX. Figs. 3, 4, 7 and 8. The color of the lavas is 1) According to Bowen and Andersen ('The Binary System MqO — SiO^', Am. Jour. Sei., 1914, p. 499), olivine crystals may be resorbed during the normal course of crystallization as a simple result of cooHng, and they inferred the formation of reaction-rims of enstatite around the crystal to the same cause. The change of forsterite to clino-enstatite is said to take place at about 1,557°C. As the oUvine sometimes appears associated with feldspar making micronoritic motex, as in the historic Salmra-jima lavas, any motion, whether upward or downward, of the crystals in magma can not be totally denied, as was demonstrated by Bowen (' Crystallization-Dififerentia- tion in Silicate Licpiid,' klein, (1915), pp. 175 to 191). In the present case, we have to do with a ferriferous complex system, and the remarkable pseudopodiform iron ore forms the transitional zone between the secondarily formed pleochroic hypersthene aggregate and the resorption-rest of an iron-rich olivine. THE GREAT ERUPTION OF SA.KURA-JIMA IN 1914. 179 blackish, and the nature of the groundmass is responsible for various shades, being entirely dependent upon the presence and quantity of magnetite grains and globulites in the base. The dark gray arises from the presence of globulite, and the intensity of darkness is caused by the quantity of magnetite, just as in synthetic gems the intensity of colors of a substance depends upon the large size of the pigment. The bluish tinge indicates the presence of light violet-brown glass ; the globulitic glass is never colored. What is stated above, chiefly refers to the main mass of the western lavas. There are, however, accidental varieties of which the writer is able to differentiate tico types. Both are found either at the lateral margin or on the terminal lava front, and character- istically free fi'om olivine, owing probably to the fact that they are the salic precursors of lava effusions, and that the olivine sank down in the main magma by gravitative differentiation. The a type (PL XIX. Figs. 3-4) is outwardly characterized in being of grayish color with various shades, and of dominantly feldspathic nature. The groundmass is the pilotaxitic feltwork of feldspar microlites with a smaller amount of augite, which is equal in size with the former. The glassy base is colorless with much or no globulite. Microscopic flecks are the spots where globulites are specially accumulated. The ß type (PI. XIX. Fig. 5) is a brown obsidian in the petro- gi'apliic sense. This liyaloandesite is a variegated, dappled and streaked glass, variously kneaded in wavy bands marked with series of axiolites and flattened air -pores. Dark flecks are seen as in the a type. All the historic lavas of Sakura-jima, including the most recent ones, have the same appearance, structure and raineralogical 180 ART. 3. — B. KOTO : composition, with the only difference that in the lavas prior to 1779 olivine is wanting. J>) Chemical Characters.— An analysis was made of a very slaggy land lava of the Nabé-yama area by Mr. S. Tanaka, of our Geological Survey, with the result given in I., besides an in- complete analysis of an upper crust of the lava sheet in the western field in II. ♦ , I. II. SIO, GO. 50 62.15 Al,0, 17.77 — Fe,0, 1.23 — FeO 5.59 4.69 31 g 0 •2.39 2.17 CnO 6.34 6.05 Na,0 3.04 2.72 IQO 1.68 1.56 H,0 0.59 0.63 TIO, 0.71 0.65 P,0, 0.08 0.12 SO, 0.23 — 3In() 0.24 0.17 100.48 The Nabe-yania land-lava. Fusing point coriesponds to No. 2 Segel cone =3 ±1, 170^0. Specific gravity at 15°C. is 2.64. The present olivine-bearing liypersthene-andesite contains large amounts of GoO, 3IgO, Na.,0, TiO., and MnO, especially the fii'st, but less of /S'iOj, ^vhen compared with a typical liypersthene- andesite. THE GREAT EllUPTlON OF 8AKU11A-.JIMA IN 1914. 181 Norms '^ Quartz 16.81 Oi-tlioclase 9.95 Mhite 24.16 Anoi-thite 30.63 8od. sulphate . . . . 0.41 Water 0.59 H^^perstheue . . . . 14.33 Diopskle 0.28 Apatite 0.12 ^Magnetite 1.79 Ilmenite 1.35 Q 16.81 F 64.74, Sal 81.55 Fern 17.87 Sal 81.55 Fern 17.87 16.81 F 64.74 K,0' + Na,0' CaO' KM' 179 = 4.57<-^>- 1 3 ^ 1 : 0.26< — > — 5 7 669 . 1132 3 1 5 7 NaM' = 0.36<— >i 490 5 7 Class n. Order 4. Ransc 4. Subrane; 4, Bandose c) Submarine Laya-flows (PL XIX. Figs. 6, 7 ; PI. XX. Figs. 2, 3 and 4).^The submarine lavas of the loestern field were taken out by a diver under the writer's supervision at three spots from depths of 10 to 40 fathoms at a distance of 20 to 80 m. from the edge of the seaward encroaching lava- flows. The writer therefore believes, that his specimens are not fragments wliich slid fi'om the advancing front of subaërial lava. Contrary to our expectation of finding pillow-lavas,'^ they form an arched-up (1 o m e - s h a p e d solid sheet of lava, externally fractured and 1) Calculated by Ogura and Matsumoto. 2) Pillow-lavas are believed to be only formed from feraic magmas. H. S. Washington, « The Submarine Eruptions of 1831 and 1891 near Pantellaria.' Am. Jour. Si'i, XXVII. Feb., 1909. 182 ART. 3. — B. KOTÙ : mncli modified in appearance from the original smooth surface. The specimens were hammered off three and a half months after the extravasation, when the lava-field above sea surface was still fuming with the s a 1 m i a c and other vapors, while the submarine portion was already partially covered with sea mosses and serpula worms. The submarine lavas that were discharged from vents (secche) in the bottom, and those (bocche) from land consolidated in the subaqueous environment, difier greatly in their mineralogical development of magma ; the former resemble an artificial sl^ig, well exemphfied in the rock of the new islets of the An-ei era (pp. 54, 176), while the latter are more like the normal rock, though both consolidated at the sea-bottom. The former are foamy brown glass, containing fibrillated augite microlites (clino-enstatite or enstatite- augite) mingling with only few or none of feldspar laths and magnetite gi-ains (PI. XIX. Figs. 6-7 ; PI. XX. Figs. 2-4), but the latter, now under consideration, are more crystalline, as in the case of the new land rocks described above. The rock is dull-grayish black and cavernous, dopatic to sempatic, dotted with the feldspar phenocrysts (2 mm., rarely 4 vun.) ; coccolithic or lumpy, due to chilling under water, result- ing in the formation of regular, sometimes diverse networks (4 mm. in diameter) of clefts, lined with limonitic iron ores and dusts. The same structure also appears in subaërial flows, when it is drastically cooled at the lateral margin, e.g., at Sambon-gachi on the western flow (p. 72, footnote 4, and Text-fig. 18b). The phenocrystic plagioclase (labradorite) is full of brown glass due to temporary retrogressive corrosion, and is zonal- structured owing to rapid growth of the interior and also renewed growth in a later phase in another environment. Microphenocrysts of olivine are almost always present (PL XIX. Fig. 7 ; PL XX. Fig. 3) THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 183 although quantitatively insignificant. They never occur in euhedral form, as we usually find them in basalts, and frequently constitute nuclei of cumulophyric patches (Text-fig. 40, p. 178). Phenocrystic pyroxenes are scarce and seriate, and deeply corroded, of wliich subhedral hypers fchene generally dominates over anhedral augite, although their proportions are sometimes reversed. Parallel growth is often observed in larger pyroxenes. Magnetite occurs sporadi- cally in clumps and crystals. All the phenocrysts are affected by frequent shifting of the equilibrium of surroundings until the final consolidation of the lavas has taken place. The groundmass of all specimens is very fine (PL XIX. Fig. 7, PL XX. Fig. 4), brownish and hyalopilitic, being mainly built up of dusted augite microlites mixed up with a subordinate quantity of skeletal feldspar laths in a light-brownish globulitic base.^^ The submarine lavas of the eastern field (PL XX. Figs. 2-4) were collected by the writer in April, 1914, at the lava-front at Séto, and 2 specimens by F. Omoki in April, 1915, the latter being from the raised heads of submarine flows now forming new rock- islets off-shore. The pétrographie characters are the same as on the aforementioned western side, excepting the presence of large olivine (8 mm.) as in PL XX. Fig. 3. The external slaggy portion of the lava- sheet is, however, wantinginolivine owing to gravita- tive sinking of that mineral, and at the same time globulitic colorless glass makes its appearance in the groundmass. In short, the submarine lavas also belong to the category of the ohvine-bearing hypersthene-andesite, and are characterized firstly. 1) H. S. Washington says, that the groundmass of the Italian submarine lavas is a clear or sometimes brown glass thickly sprinkled with rusty brown or black dust, so much so as to be almost or quite opaque in places. ' The Submarine Eruption of 1831 and 1891 near Pantellaria.' Am. Jovr. Sei., XXVII. Feb., 1909. The globulitic or thickly dusted glass seems to be the imi- versal characteristic of submarine lavas in the recent as well as in older ones in Sakiira-jima. 184 ABT. 3. — B. KOÏÔ : by coccolithic structure, secondly, by comparative idiomorphy of plagioclase, thirdly, by the presence of large sporadic magnetite clumps, and lastly, by the dusted augite microlites in the light- hrownish glass of the gi'oundmass.^^ The olivine and hypersthene have close paragenic relations ; when the former appears in noticeable quantity the latter correspondingly increases in amount. d) Characteristics of the Western and Eastern Lavas. — All the ' hve' lavas of Sakura-jima have the same characters, not only macroscopically but also microscopically, in their black color, their mineralogical composition, and also in their texture. The same holds good generally for all the historic lavas of the island. As efïusives are, in contrast to plutonics and hypabyssals, not easily susceptible to magmatic differentiation, they naturally consolidate into a uniform homogeneous mass. To find hidden, slight and constant differences among the recent lavas is, therefore, not an easy task. In the preceding, the writer has characterized (c) the sub- marine lavas from both the terrigene- subaqueous and genuine sub- marine origin and {a) the main land flows. Slight deviations in habits and also in chemical constitution may i^erhaps be observed in the lava streams of distal and proximal ends from vents. As ' hve ' lavas, while crawling in their downward course, describe comphcate comminglings and rotatory movements in a horizontal as well as in a vertical sense like streams and glaciers, specific differences, if any, in chemical and textural points cannot be easily recognized. In the foUowing, some prominent features of the lavas from the western and eastern vents will be griven in a brief form. In the bilateral eruptions of Sakura-jima, the eastern vent at 1) See ante, 54 ami 176. THE (iEEAT ERUPTION OF SAKURA-.TIMA IN 1914. 185 an altitude of 300 m. burst out 10 minutes later than the western ones, located also at 300 m. {See Text-fig. 14 b, right west, left east). The outpouring of fluent lava, which began eight hours after the first outburst, must have been correspondingly later in the eastern vents, and the latter lavas should be assumed to be of younger birth. Consequently, it is not unreasonable to suppose that the eastern flows came from a lower horizon {see Text-fig. 30) of reservoir where, prior to the eruption after long quiescence, magmatic differ- entiation might have taken place either by selective crystallization or by gravitative concentration, or else in some other w^ay. The eastern lavas on tliis account are heavy, more Ijasic and thin-fluid, flowing down farther seawards in submarine flows as compared with the western ones. {See Text-figs. 19 and 24.) Again the eastern lavas are less slaggy (comparatively compact), deeper in color and higher in specific gravity, the lustre being, however, variable in rocks on both sides ranging from pitch-black to vit- reous and dull-gray. A peculiar feature, which attracted the writer's attention, is the abundancy of lapilli on the eastern side.^^ The sea was at that time thickly covered with floating pumice fragments only on the eastern water but not the western channel. The basic eastern lava, saturated with volatile constituents, seems to have been well adapted to the formation of spongy éjecta. It is superflous to say that the west winds during the eruption blew suspending and projecting éjecta toward the east, thereby contributing a mass of subaërial fi-agmentary volcanics to the eastern side in no small measure. Phenocrysts. — Olivine,'^ which is always anhedral, is universally 1) See aate, pp. 68 (footnote), 82. '2) See tviüe, page 177. 186 ART. 3. B. KOTO : present in the eastern lavas, while in the western it appears only sporadically, especially in submarine lavas, where it is associated with amoeboid -shaped iron ore fringed with secondarily formed prismoids of hypersthene (Text-fig. 40). Pyroxenes are present in variable quantities on both sides, of which hypersthene decidedly prevails over augite in the east, while in the west both com- ponents appear almost in equal proportions. Parallel growths of pyroxenes are more frequent in the east. Plagioclase, being always zonal- structured, occurs very fi-equently in the east in segregation- patches in the form of either micronorite (plagioclase-hypersthene motex. Text-fig. 36)'^ or microtinite (pure anorthite aggregate)."^ Generally speaking, these enclaves of earlier segregation are characteristic to the ' live ' lavas of Sakura-jima. Magnetite occurs in larger clumps in the east, as compared with its idiomorphic development in the west. Geoundmass. — All are hyalopilitic ; the glass base being either brownish or colorless in the east, while the latter prevails in the west. The relative proportion of glass, augite microhtes and feldspar laths vary according to circumstances, on which entirely depend the lustre and shades of the color of the lava. Augite microlites appear in two modifications, one being clear and euhedral, the other corroded and sprinkled with dust of magnetite. The second type is more frequently met with in the glass-rich variety mingled with only a few feldspar laths. Glassy rocks are abundant in the west. The brown glassy form of ex- ceptional type is also met with in the west, which shows damascened texture, caused by complicated kneading of fl.attened air-pores and dust streams of magnetite (PL XIX. Fig. 5 ; PI. XXIII. Fig. 4), 1) See pp. 19a-194. 2) See p. 190. THE CJREAT ERUPTION OF SAKUEA-JIMA IN 1914. 187 What is stated in the preceding as the distinctive features of lavas on both sides of Sakura-jima is subject to rectification when additional shdes can be examined, especially with regard to the statement on the groundmass. 0^ B. The Loose Ejecta. Under this heading are embraced those loose volcanic éjecta - monta which are of pecuUar interest, and some of wliich are new to science, as may be learned from the following enumeration : 1) Volcanic scum.^^ 2) Gabbroids as éjecta. 3) Ceramicites, natural porcelain éjecta. 4) Solid ejected blocks of juvenile lavas. 5) Bread-crust bombs. 6) Recent lapiUi. 7) The porpliyritic obsidian. 8) Pseudobombs of ancient lavas. 9) Ejecta of trass or ash-stone. 10) The so-caUed sandstone. 11) Ejected blocks of biotite- granite. 12) Coal. 1) Yolcanic Scum.— By the term, volcanic or ' lava scum ''•* 1) On account of the presence of a large quantity of cordierite, the whole group of the scum or froth should be better consigned to that of ceramicites, which name the writer ven- tures here to proi30se. 2) It should not be confounded with the so-called ' spongy thread-lace glass scoria,' descri- bed by Dana in his " Characteristics of Volcanoes," p. 163, and by I. Friedlaender in Zeitschrift fur Vulkmiolorjie, Bd. I. Tafel XXHI. Bilder 20 und 21. The genuine thread-lace scorLa was, however, ejected from Asama in the eruption of October, 1914. The scoria is of large size, snow ■white, and hght like soap-bubbles. Microscopically, it has a peculiar sponge-like structure with .spicxiles regularly arranged forming a framework of the light mass. The same frothy scoria was blown up from the Yimo-hiia vent during the first phase of the recent eruption. It is of large size, measuring up to half a metre in diameter. The thread-lace scoria has recently been a subject of much discussion. F. A; Perret saw this Kilauean scoria lying upon the pahoehoe lava in a continuous stratum, indicating formation in situ, and therefore not to be included under the head of ejectamenta (' Some Kilauean Ejecta- menta.' Am. Jour. Sei., Vol. XXXVI. 1913, p. 617). S. Powers is of opinion that this scoria, a variety of basaltic pumice, is a kind of gaseous lava-froth blown out of craters and not formed in situ on the surface of the jjahoehoe flows. Moreover, he defines the scoria as a gaseous pumice where the vesicles are separated only by slender threads instead of walls (' Explosive Ejectamenta of Kilauea.' Am. Jour. Sei., No. 243, 1916, p. 240). The thread-lace scoria seems not to bo exclusively a kind of basaltic pumice ; for, the writer found it, as cited above, among the recent éjecta of Asama and Sakura-jima, which are built up of pyroxene-andesites. 188 ART. 3. B. KOTÔ : the writer has in mind the whitish glassy substance, consisting of aggregation of bubbles, which are formed on the surface of magma when coming in contact with the air by violent agitation of liquid lava within the mouth of the vent. The lava-froth,^^ which is a product of the Strombolian stage of activity and new to science, was projected into the atmosphere and fell on land and the sur- rounding sea. It is found on both sides of Sakura-jima, typical specimens are seen scattered abundantly on the west with its kindred ' spongy thread-lace glass scoria.' The forms are variable (Text-fig. 41) ; some being roughly flat-conoidal, and sometimes pyramidal, but most of them are irregularly rounded and always flat ; and the body is traversed Fig. 41. — Varicms forms of lava scums. Natural size. Loc. : Yokoyama. 1) A lava foam or froth, thoiigh different in kind, was noted in a retient paper by 'Tagger, Jr. (see ante, p. 93, footnote 3) as a product of ' gas-impelled and gas-heated fountains as the initial maniffistation of a rising lava column.' THE GKEAT KRUniON OF SAKURA-JLMA ]N 1914. 189 with contraction- cracks and finely textured on the periphery, while tlie inner part is coarse as in bread-crust bombs. Original shapes are greatly modified by frictional erosion while flying through the air. The size varies fi'om 2 to 5 cm., averaging 3 cm. There are 2 types in regard to colors. a) White Scum (PL XX. Fig. G). — This type is macroscopically a white, friable and saccharoidal mass, which under the microscope is speckled with polarizing particles of feldspar of 0.03 to 0.04 mm. in diameter, enclosed by elongated pores in colorless glass- film. Sometimes the feldspar particles form globular aggregates of 0.09 to 0.10 mm., radially fringed with elongated pores {0.07 mm. in longer diameters) of glassy membrane. The feldspar polarizes with brown to bluish gray of the first order. The shghtly brownish color of the scum is due to the yellowish iron globules attached inside the wall of the pores. The coarser variety is purer in color and principally made up of glass-film, enclosing larger ellipsoidal pores (0.12 mm.). A few hypersthene^^ and plagioclase crystals are present, being fringed with radially arranged and elongated pores, which are probably due to the liberation of heat during their crystallization. ß) Gray Scum (PI. XX. Fig. 7). — Although macroscopically this type, of which the large bulk of the scums consists, difïërs from the preceding in color, being ashy or bluish gray, the mic- roscopic appearance is the same in both, except the thin dispersal of finer particles of feldspar and the presence of aggregates of magnetite globules, and of very minute and round vacuoles (0.02 to 0.03 mm.). The vacuoles are responsible for producing that grayish 1) Sometimes it occurs in very fine euiiedral needles, having an appearance of brownish nitile -nith totally reflecteil dark prismatic margin. Optically positive along prismatic axis, as in Tutile. 190 ART. 3.— B. KOTÔ : effect of color by total reflection of penetrating rays of light. It is worth while to note that pores of the size of 0.07 mm. have httle effect on the microscopic appearance, while those of the minute size of 0.02 to 0.03 mm. lend an allochromatic gray color to the scmn. The minuteness of size, of course, contributes to the in- crement of vacuoles in a given space. All the scums, including the white and gray types, contain bluish cordierite in moderate quantity, sometimes filled with glo- bules of pyrrhotite, which habitually forms the centre of radiating air-pores (PI. XX. Fig. 7, PI. XXI. Fig. 1). They may, therefore, appropriately be called cordierltiferous scums, and their systematic position should be allotted to the ceramicite group. 2) The Gabbroids as Ejecta (PL XX. Fig. 8, PI. XXL Figs. 2-5, PI. XXII. Fig. 1). — The projectiles of the gabbroids of segi^egate nature are found in 4 forms. The « type is an elliptical bomb of about 10 cm., thinly coated with black slag. It is friable, easily to be crushed between the fingers, and its fracturai surface has a white saccharoidal appearance with vitreous lustre. The ß type is found in the form of flat discs. The /' type is met with in angular frag- ments in black compact andésite. It has a porcelain-like aspect (ceramicite) and resembles closely the weU-known cordierite-bearing éjecta of the volcanoes of Asama and I wate. Microscopic patches — the o type with the composition and fabric of norite, are fre- quently noticed in the recent lavas from both the western and eastern vents. Lastly, the e type occurs in enclaves, having an eucritic composition, which deserves the new naîne of micro- allivalite. T/ie o. type—'^ microtimte (PI. XX. Fig. 8, PL XXL Fig. 2), is seen under the microscope to l^e composed of polygonal crystals of anorthite, closely fitted together as in contact texture, thereby THE GREAT ERUPTION OF SAKÜRA-JIMA IN 1914. 191 causing a maeroscopically saccharoiclal effects. It is an anortJiosite- equic aient in lava. The anorthite shows no signs of plagioclase twinning, polarizes in brilHant colors with closely packed banding of spectrum colors on periphery, distincly seen in Figure 2 of Plate XXI. The size varies from 0.15 to 1.50 mm. The only other constituent is orthoclase (?) which fills, in patches {om?n.), the interspaces left by the aggregates of anorthite (the black patches in Fig. 2, PL XXI. and the dirty field in Fig. 8, PL XX.). The grayish color of the orthoclase is solely due to the multitude of pores as in the quartz of mylonitized granite. Birefringence and refraction are lower than those of anorthite. Quartz, if present, is not easily distinguished from others by ordinary method. Laceoix mentioned the same blocks ejected from Martinique and Santorin. This particular specimen of the enclave was qualitatively tested by Kanai, of the Higher Agricultural College of Kagoshima, and gave the following results : SiO-z very abundant, CaO and AIM abundant, alkalies {KiOyNa-^O) moderate, Fe little, MgO none. Presumably this white projectile is composed of anorthite mixed with a subordinate quantity of orthoclase, as is proved by micros- copic analysis. The ß type (PI. XXI. Fig. 5, PL XXII. Fig. 1) is a white disc of 10 cm, with black banding tinged slightly green (Text-fig. 42). It is rarely found and hitherto only in the east. It has a fine granular porcelain-like appearance on fracturai surface, and the mineralogical composition of cUopside-gabhro with aplitic texture. Basic plagioclase forms an equigranular aggregate of subhedrons and grains, some showing symmetric extinction at 40^, indicating that we have before us either bytownite or anorthite. Another granular component is anhedral diopside with augite-cleavage, zonally 192 AUX. 6. — B. KOTO : 'y^^h. 4;:,p colored light -green in centre, then bluish green, and lastly, brownish on margin. The last zone may be easily mistaken for a dioritic horn- blende. The last but not least com- ponent are color- less . elongated plates with high birefringence, posi- tive character and a maximum ex- tinction-angle of about 26 \ It seems to be tre- molite wdtli amphibole cleavage.'^ Tiie whole is cemented by w^hitish or dirty residual glass. The y type— the lithoid microtinite, is, in contrast to the a type, dull- white and slightly porous, and closely resembles the third type of cerainlcite (p. 200). Microscopically, it is composed of anhedral microphenocrysts (0.4 mm.) of plagioclase imbedded in the groundmass of denticulated grains (0.040 to 0.015 mm.) of the same mineral. A few slender needles of diopside are intermixed Fig. 42.— Ejecta of diopside-gabbro, a secretionary i>rodurt of angite-andesite. Eight-hand figure is the cross section. Log. : Eastern slope of the Middle Cone. 101 X 8 <™. 1) It looks very much like ■nujI.la.sfißv.Uc in the nodule comijosed of anorthite and pyroxenes, which nodule is allotted to the category of the polygciiovs endogene enrjavs in Lacroix's 'Montagne Pelée et ses éruptions,' p. 537. a'HE GREAT ERUPTION OF SAKURA-JIMA IN 1914, 193 with the mass. Tho porous portion has mor- tar texture, the round plagioclase (? cordierite) grains I)eing cemented with colorless glass full of vacuoles. Plagio- clase^^-phenocrysts are composed of an aggre- gate apparently forming trillings like those of cordierite, but their mineralogical nature is not yet definitely settled. The writer considers this angular white en- clave (Text-fig. 43) to be petrologically of diopside-gabbro. The 0 type — a micronorite (Text- figs. 32 and 36). An enclave of noritic composition and motex has been known since the pubhcation of Laceoix's work on Montagne Pelée. In Japan the writer observed the same enclave, of a few centimetres in size, among the effusives of the volcanoes of Ontaké (Prov. Shinano) and Komaga-daké (in Hokkaido). In the ' live ' lavas of Sakura-jima, patches (1 mm.) of noritic segregation are of « J) io Fig. 43.— White lithoidal microtinite in andesite-ejecta, fonnrl at Yokoyama. 12 x Sohl 1) Some of the plagioclase now under question may be a cordierite complicately multiple- twinned and penetrated after (110). Some are entirely composed of a cumulative aggregate of colorless rectangular and square-shaped embryo-crystals ^^•ith diagonal extinction. They belong in all f)robability to cordierite. 194 ART. 3. — B. KOTÙ : quite common occmTence (pp. 156, 172, 186), although they are not usually recognized as such under the microscope. Hypersthene and basic plagioclase associate either with ophitic relation, or in hypidiomorphic granular aggregate. Olivine, if present, occupies the centre of patches peripherally corroded and fringed with vermiform magnetite. See Text-figs. 39 and 40. The £ type (PI. XXL Figs. 3-4). In April, 1915, F. Omoei found amongst éjecta a round reddish granitic body, 8 cm. in diameter, coated with black vitreous lava, between the second and third vent of the eastern side. On close examination it turned out to be a eucritic mass^^ of probably segregate origin. A specimen of similar motex is in our Institute from the island of Nii-jima of the Izu group. It is an hypidiomorphic aggregate of anorthite (0.8- 1.5 mm.) and small pyroxenes {O.l^j mm.), including hypers- thene and augite in ophitic fabric. Anhedral olivine (0.15 mm. and less) is abund- antly represented, being con- stantly enclosed in the feldspar (Text-fig. 44). Anorthite is of the same character as in the aforementioned gabbroids. This stone-meteorite-like ac- Kig. 44. — Feldspar-phenocryst iu the eastern lava-field at Yébino-tsiika. x 18. 1) The rock closely resembles an aerolite called evicrlte not only in its minéralogie com- liosition, but also in its motex, being loose-granular, and the essential component, anorthite, also shows anomal double-refraction and undulatory extinction. Being absolutely holocrystalline hyp- idiomorphic and partly iJoikiUtic, we fail to find interstitial glass cement of a composition of feldspar called maskelynite in ' Meteoritenkunde.' Here the interstitial spaces, if any, are void and emi)ty. THE GKEAT ERUPTION OF S.VKURA-JIMA IN 1914. 195 cessoiy mass of rock deserves the new name of micro-allivalite, since it is mainly composed of anorthite and olivine, corresponding to micronorite-patches in andésites first described and named by Laceoix in his well-known Martinique work. A similar eucritic intergrowtli or rather enclosure of anhedral olivines in a basic plagioclase is to be seen in the feldspar-pheno- cryst of the recent lava near Yébino-tsuka hill in the eastern lava- field, as in Text-fig. 44. The whitish masses dealt with under the present heading are usually interpreted as formed under p 1 u t o n i c condition, and caught up by magma and brought to the surface. From the mortar texture with glass cement, pumiceous texture and pseudophenocrystic (cumuloporphyric) occurrences of masses, it seems they may not necessarily be formed in a deep horizon, and especially the noritic patches are likely formed even during subaërial flows.^^ Anyway they are products of relatively early separation, and composed of minerals of high melting point among andesitic components. Selective crystalhzation and resistance toward resorption of the components caused magmatic differentiation, and concentrated the minerals in certain spots. The segregated portions were then conveyed through upward current to the surface. They are not, however, necessarily buoyed up in viscous magma from a deeper horizon, where the minerals gather owing to their comparatively high specific gravity. It is to be noted that the enclaves of diopside-gabbro are 1) The macroscopic banding, e.r/., as in the ß type (p. 191), is due to special aggregation of diopside and tremolite in one plane. To the writer this si^ecial texture seems to suggest the formation of the mass near the surface or on the surface of the lava-vent. Under plutonic conditions the same mass will crystalhze in homogeneous mixture, or else the banding must be attributed to primary flow of segregation — a fact which is not observalile in microscopic analysis in this granular aggregate. 196 ART. 3. B. KOTO : found in large mass, while those of noritic composition are small and intimately mixed in lava-flows, as if forming an integral por- tion of rock components. Hypersthene as compared with diopside has lower melting point and density, and these seem to the writer to be the cause of the later origin and higher level of separation of noritic segregation. Another lesson to be learned is, that basic diorite- schists and quartzless gabbros, both being usually considered to be plutonics, may be formed as accessory portions of modern effusives in a comparatively shallow horizon. The only difference observable in comparison with plutonic equivalents is the absence of garnet, which is the geologic barometer, occurring only in deep-seated rocks. It is worth while to note in connection with microtinite that the normal chemical composition of the base, or groundmass of rocks of intermediate acidity like andésites, as observed by Teall,^^ greatly resembles that of basic plagioclase. A slight addition of iron to the plagioclase-melt will result in the formation of chemical combinations of the enclaves which are mainly composed of glassy plagioclase or microtine already referred to. From a chemical standpoint there is therefore no coercive ground in the way of explanation on the formation of the microtinite enclaves in our lavas from molten magma. Similar inclusions in sills have been already discussed by some British petrologists."^ 3) The Ceramicites, Natural Porcelain Ejecta (PL XXII. Fig. 2-8). — By the cordierite- bearing éjecta are hero meant those porcelain-hke projectiles from volcanoes, which contain cordierite as the charac- 1) ' British Petrography,' p. 42-43. See postea, p. 203. 2a) J. A. Smythe, ' On Some Inckisions in the Great Whin Sill of Northumberland.' Geol. Mag., 1914, p. 244. 2'j) Henslop and Smythe, 'The Dyke at Crookdene.' Q.J.G.S., London, 1910, p. 7. THE GEEAT ERUPTION OF SAKUKA-JIMA IN 1914. 197 teristic component, the remaining ingredients being basic plagio- clase and colorless glass, with a subordinate amount of hyper- sthene.^^ There are gradual transitions frori the preceding gabbroid éjecta (pp. 190-196) and lava scums (pp. 187-190) to the one under question, and a sharp line of demarcation cannot be drawn between them. The only difference is the presence of a charac- teristic, magnésie, orthorhombic mineral — the cordierite, which varies quantitatively within wide range. The more crystalline the mass, the less the amount of the cordierite, which sustains reciprocal proportion to basic plagioclase.^-' If obsidian is a natural glass, the éjecta may fittingly be called natural earthemoare or ceramicite"'^ with modifications corresponding to pottery, stoneware and por- celain. Ceramic wares are characterized by the presence of quartz and sillimanite, wliile the natural products contain cordierite and plagioclase in lieu of the two minerals. The cordierite-bearing éjecta from Asama-yama are now well- known through the work of Hussak.*^ Since then other occurrences were added to the list from Gôro-yama near the city of Nagano (Zenkôji),^^ Ganju-san and Komaga-také near by, on the west of the city of Morioka,''^ and the Komaga-také of Hokkaido."^ Exactly the same éjecta were shown the winter by Laceoix,*^^ who gave a description of them. Mention may be made here of projectiles 1) Fine needles of this mineral often apj^ear under the microscope like rutile. Cf. p. 189, footnote. 2) From the measurement of maximum symmetric extinction, oligoclase seems to be pre- sent among phenocrystic round crystals. 3) The writer ventures to give this 7ieio name to the multifarious cordierite-bearing éjecta. See ante, p. 187, footnote (1). 4) 'Ueber den Cordierit in vulkanischen Auswürflingen.' Sitzungsber. d. K.K, Akad. d. Wiss. in Wien. Math.-Naturwiss. Klasse, 1883, LXXX^TI. S. 332. 5) N. Yamasaki, Jour. Geol. Soc. Tokyo, Vol. II. p. 397. Ihid. Vol. IV. p. 35. (-Japanese) 6) Found in the collection of the late Mr. Sakuiai, deposited in our Museum. 7) Tak. Katô, ' Ueber die Kordieritführenden Einschlüsse in der Lava aus dem Vulkan Komaga-také auf Hokkaido, Japan.' Jour. Geol. Soc. Tokyo, Vol. XIX. p. 27. 8) ' La Montagne Pelée et ses éruptions.' 198 ART. 3. — B. KOTÔ : from the Laacher See.^^ The garnet-bearing plagiohparitic efFusives containing the same mineral are made known from Futakami- yama, Prov. Yamato,'^ and Obora-yama,"^^ Prov. Isé, both being probably the same kind of rocks as those from Campiglia Mari- tima/^ Lipari,^^ Cabo di Gata,"^ and Tian-shan/^ The éjecta fi^om Sakm^a-jima may be divided into 5 types, according to granularity. Trie first type (PL XXII. Fig. 2) — the saccharoidal modification — is pure white and saccharoidal with blue flecks (4 imn.) of short- prismatic cordierite and dark, irregular spots (1 mm.) of pyrrhotite. It crumbles easily between the fingers like the saccharoidal micro- tinite already mentioned (p. 190, the « type). Microscopically, cordierite is porphyritic and anhedral or sub- hedral. It is peripherally resolved into subr octangular microcrystals arranged parallel to the c-axis of main crystals (PI. XXII. Fig. 3), sometimes, however, being shghtly divergent (PI. XXII. Fig. 4) like digitate crystals or ' fingered aggregate ' of andalusite in contact rocks, indicating that the larger cordierites were being formed during the turbulent state of the magma within the vent or in viscous lava. It is negative and bluish, parallel to c, and charac- teristically contains pores in the centre (PI. XXII. Fig. 2, c). In other cases the cordierite is entirely built up of minute grains of the same mineral (PI. XXII. Fig. 4). Irregular clumps of pyrrhotite of 1) R. Brauns, ' Die kristallinen Schiefer des Laacher Seegebietes und ihre UmbUduug zu Sanidinit.' 2) M. Üyu, ' Report on the A^'olcano Fiitagami-yama.' Puhlicatlons Imp. Earthq. Invest. Com., No. 27, Tokyo. (Japanese) 3) Found during revising a student's slides. 4) Zeitschr. d, deutch. 11.86 Alsbachnse K,0' + Na.O' F Columbare 46.47 ^_6L=4.79<1>1 14 13 40.31 K.,0' = 1.15< — > — 3 5 Tehamose 28 0.77< — >-?^ 3 5 CaO' 14 ^ 1 ' 3 Na,0' 39 The ceramicite is a scale segregate of augite -andésite, and chemically speaking, it represents a series of biotite-granite, quartz- porphyry, rhyolite and quartz-porphyry schist. The fleckless barren modification (PL XXII. Fig. 5, and Text-fig. 45) contains idiomorphic phenocrysts (1 mm.) of polysynthetic anorthite imbedded in the vacuole-rich glass ground, which contains a few corroded non-pleochroic cordierite (the rather large subangular 1) Calculated by T. Ogiira. 202 ART. 3. — B. KOTÔ: white flecks in Fig. 5) and minute fragmentary plagioclase. The whole microscopic aspect presents that of a tuff cemented with glass. This feldspathic modification may be called the ' faience ' of natural earthenware with fine semitranslucent texture. The fourth tijps (PI. XXII. Fig. 6).— It is marble- white and fine, light and friable. It shows banding, caused by alternation of a bluish compact zone (a) with a lustreless vesicular one {h). This projectile is astonishingly rich in ill-defined granulated cordierite (c in Fig. 6). The banding is, as seems to the writer, an in- dication that the ejected fragment in question is not segregated under plutonic conditions, nor a resorption-rest of an exotic fragment, as it is usuaJly interpreted, but a primary segregation- product of lavas near the surface. To support the writer's case he has in his possession a liluish pumice full of cordierite, ejected from Asama on December 14th, 1912. See PL XXI. Fig. 1. Microscopically, the [a) zone is exactly the same as the third, lithoidal type, while the {h) zone represents the lava scum.^^ Again, we have here an indication that the ceramicite and lava- scum are genetically closely related. The banding points to being a result of circulation of lava in the intercrateral area."^ The fifth type (PI. XXII. Fig. 7).— The wet-gray, striped, re- sinous, porcelain-jaspilite variety, breaking with subconchoidal fi^acture. The light colored stripe (a) is composed of subrectangular^^ or roundish^-' crystals of cordierite (0.025 mm.) besides a few larger ones, intermixed with prismoids of plagioclase and fine needles of hyper- sthene ; the whole is imbedded in a colorless base, which encloses 1) PI. XX. Fig. 6 shows the lava-scum zone (h) of the éjecta. 2) F. A. Perret ; Amer. Jour. Sei, 1913, p. 3i5. 3) See the photomicrograph, PI. XXII. Fig. 7 (groundmass). 4) It is Lncroix's ' colloidal cordierite.' Op. cit. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 203 vacuoles and brownish glass-drops, the former appearing black by total reflection. The dark-gray band (h) does not materially differ from the preceding (a), excepting the presence of larger patches of cordierite, irregularly distributed in glass full of vacuoles and brown amorphous enclosures, which both lend to the rock the grayish tinge. In summarizing wiiat is stated above, the cordierite-bearing porcelain-hke éjecta, here called ceramicites, greatly resembles earthenwares with various modifications resulting chiefly from the difference of temperature at which the industrial products are artificially melted and manufactured. These volcanic éjecta compare well on one hand with the crystalline microtinite (p. 190) and on the other with the cryptocrystalline lava-scums (p. 187), al- ready briefly mentioned ; they are genetically related to one another and aU three consolidated under comparatively low pressure, probably near the open vent, as in the recent pumice of Asamayama. The components found in the éjecta are anorthite, sanidine, cordierite, pyrrhotite, magnetite, hypersthene and siUimanite, besides spongy glass ; the first two scarcely to be distinguished on account of the absence of twinning structure. Noteworthy features are the absence of quartz, muscovite and even biotite, the latter being chemically akin to cordierite. The quartz is always mentioned by the writers on cordierite-bearing éjecta as making a large bulk — an error arising from misinterpretation of cordierite and feldspars. There is little probability for the presence of a large amount of quartz, the siHca being represented in the colorless glass. The pyrrhotite mentioned above is the characteristic com- ponent of ceramicite, by its presence alone the writer is able to recognize the éjecta from the rest of the projectiles. The colors 204 ART. 3. — E. KOTO : by reflected light can afford a clue in discriminating the sulphide from pyrite. Moreover, the presence of pyrrhotite is the sign of eruptivity, while pyrogenic pyrite is a product of post-volcanic activity. A rock similar in minéralogie composition to the well- known pyrrhotite-bearing cor dierite- gneiss of plutonic origin found at Bodenmais, Bavaria, is, therefore, represented in the recent lava of Sakura-jima. From the mode of the constant association of pyrrhotite and cordierite, the writer surmises a peculiar relation between them. In the grayish porcelaneous éjecta thrown up in a temporary eruption of xisama on the 13th of December, 1912, we find grayish blue stripes, which under the microscope (PI. XXII. Fig. 8) are seen to consist of an aggregate of cordierite rectangles cemented by pyrrhotite — a mass serving for the matrix of the resorption-relic of the microphenocryst of a basic plagioclase (an). The cementing pyrrhotite sustains the same relation to the enclosed cordierite microcrystals, as the xenomorphic chalcopyrite to the idiomor- phic ironpyrite of high crystallinity in ordinary orebodies. Both the pyrrhotite and cordierite are primary secretionary products of a basic andésite magma, and any explanation on the close associa- tion of iron and copper pyrites will help in the elucidation of the genetic relation of the two minerals in our ceramicites. The neovolcanic cordierite is associated with andésites and liparites, but not with basalt, so far as the writer is aware of. The cordierite-bearing éjecta are direct oflsprings of the former, wliile in the latter (plagioliparite) they occur as cognate xenoliths (Haeker). Exhaustive study of cordierite is out of place here. The writer can only say that idiomorpliic crystals, save those rectangular microcrystals, are not observed. Trillings, whose formation is sup- THE GKEAT ERUPTION OE SAKUEA-JIMA IN 1914. 205 posed to be attributable to high temperature (A. von Las aulx), are frequently seen. Optically positive, and pleochroic only in larger individuals ; c >(i = a. A characteristic feature is the accumulation of variously ori- entated grains (Laceoix's colloidal cordierite), of which the whole interior of the cordierite is built up (PL XXII. Figs. 4 and 6), and such composite crystals appear macroscopically dark-blue. When clear crystals are seen, minute pores are characteristically found in the central part, thereby coloring allochromatically grayish- b 1 u e ,^^ and the vacuoles are probably, originally, open spaces left between granules, as in porcelain pores. Larger cordierite is habitually resolved peripherally into grains (colloidal cordierite), or microcrystals rudely arranged parallel to the c-axis (PI. XXII. Fig. 3). Both the external and internal granula- tions seem to have risen from regressive resorption rather than from imperfect growth. Where roundish clear plagioclase is en- closed poikihtically within the vacuole-rich cordierite, the former is fringed with isotropic substance, as if it were a reactional rim (the corona of French authors), which sometimes contains gas pores (PI. XXII. Fig. 2). It may arise from remelting through a temporary rise of temperature. Glass makes up the colorless groundmass in variable quantity, in which an excess of SiO., found in the analysis, is likely to be contained. It is always fuU of pores. When the vacuoles sink to 0.02 to 0.03 mm., the glass appears macroscopically gray on account of total reflection of light ; otherwise it looks dull-white. As to the mode of formation of ceramicites or cordieritiferous éjecta, the petrologists are not in accord in their opinions, but 1) The writer conjectures that the characteristic hluisJi colors and pleochroism of cordierite are attributable to the same cause. 206 AET. 3. — B. KOTO : they are right in special cases, as the éjecta may not be of a single origin. Lackoix and Bkauns {Ioc. cit.) have already given a summary of views on the modes of formation, which may be grouped into the following headings : i. Intratelluric concretion of a trachytic or sye^nitic magma (Hussak). ii. Contact-metamorphic product of the caught-up schists or sedimentaries (Bkuhns and Yamasaki). iii. Crystallized mass ù'om syntectic magma under plutonic conditions (Tak. Kato). iv. Imperfectly resorbed ancient rocks (Lacroix and Beauns). Laceoix^^ made a close study of the éjecta from Pelée and La Soufrière, which are akin to those of Sakura-jima. He is inclined, to believe the cordierite to be mostly of secondary nature, formed either in andésites or dacite magma, and recognized three possible cases of its formation. a) Eesorption-relic of a rock-fragment, e.g., granite. b) A mineral newly crystalhzed from the melt of cordierite- bearing rock-fragment. c) Newly crystallized in a cordierite -free rock, which suffered chemical reaction of new magma in coming in contact with it. Around the two volcanoes, Montagne Pelée and La Soufrière, there are, according to Lacroix, diverse cordieritiferous blocks derived from preexisting unknown rocks, viz:, «) Blocks composed of zonal feldspar (oligoclase-anortliite), hypersthene, biotite, cemented with grains of quartz and some glass. Miarolitic spaces are filled with cordierite. 1) ' La cordierite dans les produits értiptifs de la IMontagne Pelée et de la Soufrière do Saint- Vincent; Comiit. reivl. 137, p. 145-147, 1903. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 207 ß) Wliite porcelain-like masses flecked witli cordierite, being a breccia of an old andésite. Cordierite is formed by tlie reaction of constituents of new andesitic magma upon a metasilicate of the preexisting one. y) Fragments of breccia cemented with opal-like matter, the fragments themselves being aggregates of cordierite crystals. Taking into consideration wdiat has been actually observed of the éjecta and what others have said of similar projectiles, the writer is forced to the following conclusions from the data given below : 1) The cordieritiferous éjecta or ceramicites greatly resemble ceramic w^ares. 2) They contain primary glass-enclosures and air-pores.^-* 3) There are no biotite nor amphibole (with the exception of tremolite), both only crystallize under intratelluric en- vironment, the cordierite being the vicariate of biotite. 4) Quartz, so frequently spoken of as making a large bulk, seems to ])e for a greater part due to misinterpretation of either cordierite, feldspar or some other colorless minerals. 5) The small, grayish, dough-shaped lava-foams, in which cordierite and feldspars are frequently found, make gradual transitions to the solid éjecta or ceramicites. 6) The minerals composing the ceramicites are of light chemical elements, and colorless, excepting a subordinate quantity of fcTriferous minerals, viz., hypersthene and diop- side. All in all, the wTiter is disposed to think that the cordieriti- ferous éjecta of Sakura-jima are of primary leucolithic consolida- 1) The presence of gas-bubbles is the sign of melts at low pressure, and the gas contained may be the one originally caught up between subcrystals, or occluded from air. 208 ART. 3.— B. KOTÔ: tion and local collection of crystals from melts under low pressure at or near the surface within the vent, where the temperature is lower than the interior, that is to say, a subaërial product, during the Strombolian activity (Pereet) ; for this reason we have lava- scum and cordierite -bearing pumice. Chemically speaking, they seem to correspond to the ground ma s s (p. 196)^^ of andésite, the phenocrysts of early crystallization sink down to a lower liorizon, thereby producing the total effect of selective crystallization- differ- entiation. The crystalline éjecta, e.g., ordinary ceramicites, on the other hand, are the assemblage or secretion of chemically light crystals in a rather lower horizon, buoyed up through the magma to the surface, assisted by the ascent of gas and the circulation of liquid lava within the vent.'^ It is difficult to say anything about the origin of the magma, whether primary or syntectic, although the former is for the greater part higlily probable. The ceramicites are consolidated leucocrates of andesitic magma, formed close to the wall of the vent at a shallow horizon. The volcanic scums (p. 187), the gabbroid éjecta (p. 190) and the ceramicites all seem genetically of the same origin, differing only specifically. They probably correspond to Lacroix' s enclaves homœgènes or Haeker's cognate xenoliths. 4) Solid Ejected Blocks of Juvenile Lava.— The ' live ' lava makes 1) "Within the exogenous contact zone of the oligoclase-granite in the Orijiirvi region, Pentti Eskola mentions a greater number of cordierite-gneisses, skarns and ores, which are said to be formed by pnenmatolytic metamorphism under photonic condition. It will, however, apj)ear at first sight paradoxical to tind their analogous prodiicts among the juvenile recent éjecta of Sakura-jima. It is w^orthy of note here, that he often speaks of the conversion of plagloclases into cordierite, which is expressed in the term of chemical formiilas. The present writer is also convinced of the fact, that there exists an intimate relation between feldspars and cordierite, the latter often poikilitically enclosing the former with the corona of glass between them. See PI. XXn. Fig. 2. ' On the Petrology of the Orijiirvi Region in Southwestern Finland.' Bull. Com. Géol. de Finlande, No. 40, 1914, pp. 167-262. 2) Perret, ' The Circulation System in Halemaumau in Lava Lake during the Summer of 1911.' Amer. Jour. Sei., 1913, p. 345. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 209 up undoubtedly a largo portion of éjecta, besides gray lapilli, and the size differs within wide limits from 2 m. to 2 cm., depending on the distance from vents and the fury of eruptions, those of later phases being smaller in size and compact. They are all black and porphyritic with plagioclase phenocrysts. Textures vary from compact to vesicular and even slaggy, lustres also from pitch-black to dull. Pétrographie characters are the same as the main flows, of which qualitative microscopic analyses and descriptions were already given (pp. 176-181, 184-187). As has been pointed out already, the lavas of the acidic, western, and the basic, eastern flow differ only in slight degrees. The éjecta of the east side are, however, rich in basic plagioclase (labradorite, symmetric extinction 30^-33°), and in hypersthene, augite being present only in subordinate quantity, and the glass of the groundmass white. Those of the ice stern side contain a large quantity of higlily biréfringent non-pleochroic augite. The nature of plagioclase is apparently the same. The prevailing type of the groundmass con- tains brown glass and augite microlites in large quantities, in contrast to the eastern flows. 5) Bread-crust Bombs.^^— The bread-crust bombs or turtle-back bombs are the characteristic projectiles of the west, varying in size fi'om 10 cm. to 1 m., and always angular in form. They are a common type all over the globe, frequently depicted and described among the projectiles of recent volcanoes. Larger ones make hollows in dashing against the ground, whereby the projectiles themselves break open, exposing the gray spongy interior, the color being solely caused by fine inclusions of gas. The exterior is black and compact, passing by degrees to the brown and gray 1) On the formation of bread-crt;st bombs see p. 114. 210 ART. 3. B. KOTJ : interior, and traversed with cracks which usually radiate from deeper clefts of the interior. They are the juvenile lava half consolidated within the vents, where they burst open and were thrown up into the sky. There the outer portion is farther cooled down, while the in- terior is vesiculated and expanded by relief of pressure, thereby creating cracks in the glassy skin. The lapilli, abundantly scattered about, seem to be mainly fragments of the inner portion, rent, open through mutual collision of the bombs while flying in the air. Tlieir pétrographie nature is hypersthene- andésite with hyalopilitic texture. The brown glass of the groundmass swarms with equal quantities of augite and feldspar microlites. Microscopically, the rock is exactly the same as that of the lapilli, except the textare, it is therefore not necessary to restate here a pétrographie description (p. 211). The following points are, however, wortliy of note : a) The flowage texture of dopatic hyalopilitic groundmass has no relation to the surface of bombs. h) (Quantitative relation and the distribution of phenocrysts are uniform throughout the mass. c) Microlithic bodies are quantitatively uniformly distributed in different portions, contrary to expectation. d) All the pétrographie elements were already there during the aerial journey, the crystallinity of bombs having already been attained during their intercrateral stage. e) The only special aspect to be noted is the vesiculation of the groundmass, wMch causes macroscopic differences in color, texture and lustre. ./') Yesicula3 during their formation shifted microlites away sidewards. THE GREAT ERUriTON OE SAKUR.V-JTMA IN 1014. 211 (j) YcsiGuisci arc loss in number around larger vesicles. h) .Vround the hollows, microlites are tangcntially ai'ranged parallel to the margin, indicating that the lava still re- mained in a semifluid state during the gaseous expansion. /) The sudden relief of pressure is solely responsible for the evolution and coalescence of gas pores. 0) Recent Lapilli ('little stones').-/-/) Lapilh or pumiceous blocks making a large Inilk of éjecta around the iccsteru vents are externally rusty brown and internally fresli and ash-gray, dotted with plagioclase crystals (2-3 mm.). The feldspar shows a (/laij fracturai surface streaked with polysynthetic lamellar sutures, and the gray color is due to botli brown glass- and granular en- closures. Pleochroic hypersthene and dee\) greenish-brown augite (hypersthene>augite) are mafic phenocrysts. Micronoritic patches of hypersthene and bytownite, and gabbroic ones were noticed. The pumiceous hyalopilitic groundmass is of hroicnish glass, in which are scattered fibrillated augite-needles, like spicules in sponges.^^ A subordinate quantity of feldspar- laths is intermixed with augite. The brownish outer portion of lapilli contains a globulitic body in the vesiculae through oxidation in contact with air. ß) In contrast to the lapilli of western Sakura-jima mentioned above, tliose which fell in the Ushiné coast of Osumi Province, which were thrown up from the eastern vents, are of lighter shade (PL XXIII. Fig. 5). The microscope reveals not a few' dis- tinctive features. Augite is apparently lacking, and hyperstliene represents tlie whole pyroxene."^ Plagioclase is more basic, iiighly biréfringent and more distinctly zonal- structured. Brownisli an- 1) It is analogous to the structure of the thread-lace scoria (p. 187, footnote 2). 2) PI. XXIII. Fig. 6. Isolated by washing after crushing of the lapilli that fell on Tabé in the l'shintî coast near the defunct strait of Séto. 212 ART. 3. — B. KOTÔ : liodral olivine is accessorily found. The dominant groundmass is colorless pure glass, and pniniceous. In short, the eastern lapiUi are of comparatively basic nature, corresponding to the massive lava-flows, as has been already empha- sized elsewhere (p. 184). 7) Porphyritic Obsidian (PL XXIII. Fig. 4). — The single specimen of porphyritic obsidian found was picked up from among the c^jecta on the east side at an altitude of 400 m. It is a compact pitch- black vitreous glass dotted with grayish lustreless specks of plagio- clase. Slides show abundant and seriate plienocrysts of plagioclase and a subordinate amount of pyroxenes (augite>hypersthene), besides magnetite clumps in a brownish damascened glass, — a texture re- sulting from kneaded and compressed air-pores as in the figure. In a certain homogeneous portion of brown glass, brownish axiolites and colorless crystallites are found in fluidal arrangement. A substance similar in texture to obsidian occurs as coating of trass (PL XXIII. Fig. 3, and Text-fig. 4G, p. 215) and other projectiles. 8) Pseudobombs of Ancient Lavas.— There are éjecta, though quantitatively insignificant, of ancient lavas of various shape and size, more or less rounded through subaërial mutual abrasion, which may appropriately be called pseudohomhs. Their common feature is the ash-gray color ; but the texture varies from compact and tabular to somewhat scoriacoous. «) One kind consists of hypci^sthcne- andésites, and their petrogra- pliic character is the same as the Kita-daké type, to which the reader is referred for details (pp. 155-158). Specimens were only picked up on the western slope. During the earlier explosive phase vents were rent through the rock of which the western slope is built, and the ruptured resurgent fi^agments were thrown up among the juvenile lavas. THE GREAT EKUmON OF SAKURA-JDLA IN 1914. 213 ß) Among the gray ancient rocks, there is a light-gray rongh trachyte -looking éjecta. Under the microscope, weakly pleochroic, old-looking hypersthene is enclosed by a peripheral zone of brownish ' hornblende.' Gray apatite and tridymite were also observed. This hornhlende-hypersthene andésite might represent the mesa sheet above the lapilli lied of the Plateau Formation (Younger Tertiary or Early Diluvium) of southern Kyûshû {Ax in Text-fig. 3, p. 14). If the writer's supposition proves to be right, the foundation of the volcano must bo assumed to be the Plateau Formation, especially in taking into account the associated lapilli, which have been described as the 'éjecta of trass.' >S^er^ page 214. /') Among the éjecta of the western slope, an angular grayish enclave of secretionary origin was collected, which was caught up in a mass of the Furuhata lava or the North Cone type. This ' homogenous ' inclusion (PL XX. Fig. 5) has microlithic texture, being built up of microlites of feldspar (1 mm. in length), divergently arranged with polygonal spaces left between them, as if residual glass leached out through a net-work of feldspar- skeletons. The microscope reveals that augite crystallized ophitically with feldspar- microlites, and the thin membrane of colorless glass stretches between microlithic webs. This is called the ' structure enchevê- trée ' (halter strap structure) by Laceoix.^^ o) There are some round blocks with black coating. They are the disrupted fragments (Kita-daké type) of the mountain body, that fell within the caldron of broiling lava and were again thrown out in the proximity of the ventholes. The effect produced in 1) ' La Montagne Pelée et ses énijations.' PI. XXX. Fig. 6. From Dr. Sidney Powers, the writer received a specimen of olivine gabbro, taken from the walls of Kilauea, described by R. A. Daly as a wehrlitic intriisive arching np ash bed after the manner of laccolith.- — ' Igneous Rocks and Their Origin,' p. 76. On seeing a slide of it iinder the microscope, it shows the structure enchevêtrée as in our secretionnry patches, although oli\'ine is wanting in the Saknra- jima rock. 214 AKT. 3. — B. KOTÔ : the lava bath is simply caustic and slight, showing that the tempera- ture in the vents was far below 900^C. Augite-microlites, as w^ell as phenocrysts, are hematitized rusty brown, and the coloi*less glass has become granular. The alterations occur patcliwise near the contact. 9) Ejecta of Trass or Ash- stone {Uai-ishi). PI. XXIII. Fig. 3. — There are peculiar éjecta, though quantitatively insignificant, among the projectiles in western Sakura-jima. They grade in color from brick-red to brown and finally to gray, and are trass-like poly- miktic in composition, being built up of fragments (4 mm. to 1 cm.) of the black Mesozoic slate and black porphyritic augite-andesite, whitish pumice and grayish hornblende-biotite-hypersthene andésite. The agglomeratic mass is sometimes coated witli a black rind of recent lavas (Text-fig. 46). The éjecta may be grouped into two forms : «) Grayish halh of more than C cm. in diameter, without any coating of black rind. The form is the result of frictional abrasion of friable projectiles while flying in the air. The balls or pseudo- bombs effervesce with acids. ß) Angular fragments of various sizes, sometime attaining the considerable dimension of 1 rn., and in one or two instances making- hollows in the ground by projectile impact. The fragments, oi' rather blocks, are always coated partially with black rind as in Text-fig. 4G, and non-effervescent with acids. The core in contact with the rind is colored brick-red through oxidation, gradually changing into brown and then gray. Under the microscope, the black vitreous white-spotted rind, 5 cm. thick, proves to be a vitro-pyroxene andésite (porphyritic obsidian) with, the phenocrysts of plagioclase, hypersthene prisms and augite grains (hypersthene >augite). The bulk of the rind is made up of brown glass, which is crowded with skeletal microlites THE GKEAT ERUPTION OF SAKUßA-JIMA IN 1914. 2U Fig. 46. — Trass, a magma-soaked lapiUi aggregate, partially coated with obsidian rind. Loc. : Koike. of plagioclase mixed with a subordinate quantity of those of augite. It may perhaps have consoHdated from partially resorbed magmjt from trass mass, i.e., chemically speaking, syntectic. The trass core contains fragments of the underlying Mesozoic slate. The second component is dark brown augite-andesite. The third is gray, hornblende -bearing biotite-hypersthone andésite or ash-stone (hai-ishi). The grass-green hornblende and weakly pleochroic prismatic light-colored hypersthene are the mafic ele- ments, besides accessory biotite and diopsidic augite and abundant plagioclase (andesine) phenocrysts. The groundmass is higlily feldspathic and full of feldspar microlites in colorless glass. The component rock is an ancient piperno andésite that built, up the base of the volcano. The fourth and the last component is the pumice derived from the preceding, and the pumiceous hollows are entirely filled with calcite. The third and fourth together make up a large bulk of the trass. The piperno^-" andésite and pumice, as already stated, built up the extensive plateau of southern Kyûshii,^^ and constitute the 1) See page 13 and footnote on hai-ishi 2) See page 19. 216 ART. 3. — B. KOTü : foundation of the volcano. They were burst and thrown out fi'om the vent, commingled with other éjecta. That they are ancient ultra- Sakura-jima andésite is proved by the presence of hornblende and hiotite in fragments. The volcanics of above characters are entirely wanting in the indigenous lavas that compose the body of Sakura-jima. The infilhng of calcite in pumiceous space is worthy of note, and it is not found in the trass with lava rind, as calcite dissociates at 900°C. under normal pressure. The calcite is at times vicariated by a brownish isotropic irregularly-split ' opal '-hke substance, which greatly resembles imperfectly fused feldspar grains in pottery. 10) The so-called Sandstone. PL XXIII. Fig. 2. — Among the western éjecta, the rock-collectors find grayish banded fragments called sandstone in Sakura-jima. On the banding plane glimmer- ing white flecks of biotite are seen looking like muscovite- sandstone. The fi'acture at right-angles to the former shows wet-gray resinous lustre. It is the pumiceous hypersthene trachyandesite bearing hornblende and biotite {ho in Fig.), and constitutes the founda- tion of southern Kyûshù and also the base of the insular volcano. The two minerals are well seen in the photomicrograph, the hornblende being darkened {ho) through the caustic action of recent magma. Under the microscope, the resinous lustre is seen to be caused by infilling of dirty amorphous silicate in the pumiceous hollows. This glass- soaked rock was rent and thrown up like the preceding trass to the surface, intermixed with polygenous projectiles. 11) Ejected Blocks of Blotite-granite. PL XXIII. Fig. 1.— There is a unique projectile among the éjecta collected by Kanai on the margin of the main western crater, and this is a block of granite. The rock in situ can only be found at a distance of Ibhn. THE GEEAT ERUPTION OF SAKÜRA-JIMA IN 1914. 217 southeast on the mainland in Prov. Osunii, where the so-called Mesozoic slate occurs, intruded by a granite boss which bears a low grade copper ore-deposit. As stated elsewhere, Kagoshima Bay can be best explained by an assumption of meridional fault - block depression (pp. 18-20), and the basement upon which Sakura- jima rests should be identical in geologic composition to the neighboring plateau-land. The Plateau Formation (p. 14) is com- posed of a lapilli bed underlaid by marine tuffite and trachyandesite, which in turn lies upon eroded edges of Mesozoics intruded by granite. The first and second are represented among the éjecta by the 'éjecta of trass' (p. 214), and the first also by the so-called sandstone (p. 216). In addition to these, our new find of granite among the recent éjecta is of great significance in deciphering • the composition of the foundation of the volcano. Anyway the frag- ment was disrupted from the bottom at least 1,000 7?^. below^ and without being assimilated in the lava bath, was conveyed to the surface. The granite block, 10 cm. in size, is rounded on the edges by corrosion, and colored slightly brown. It is dull white and friable through fritting by hot magma, and has an impaired and faded aspect. See PL XXIII. Fig. 1. The feldspars have lost their lustre and the biotite is changed into a brick-red mass. The microscope shows the biotite to be impregnated along the cleavage plane with yellow crystals of sulpiride ore (pyrrhotite), and newly formed minute muscovite-like flakes take the place of the original biotite, some- times arranged divergently, intermixed with hematite, the fresh substance being no more seen, and appearing black in Fig. 1. The quartz is milky-white. It is minutely cracked, as if quenched, and filled with muscovite (?) flakes. Some portion is 218 AET. 3. B. KOTO : remarkably changed into an aggregate of confused straiglitly extinguisliing fibres, which may be an allotropie quartz, and the normal portion is swarmed with liquid inclosures, which are ar- ranged without relation to cracks. The orthoclase is entirely altered into an aggregate of doubly refracting pseudophitic substance presenting a blurred aspect in the figure. In short, the changes which the granite suffered, are of a caustic nature and the result of intense baking. Continuous regular long ci'acks in quartz intersecting at obtuse angles may be traceable to the rhombohedral cleavage. The sericitization of orthoclase started from twinning sutures and also fi'om externally-bounding faces, and this alteration seems to have taken place not in a low temperature, as it is usually assumed. We find no tridymite which may be thought to have originated from /3-quartz. It is stated^^ that above 870^C /^-quartz inverts to tridymite ; but unless a flux is present this change does not occur until a temperature of 1,400^C is attained. Our granite-block must have been dropped into a bath of andésite- magma having a temperature of above l,loO°C (p. 112), and yet tridymite was not formed. The friability of the granite éjecta may be attributable to the molecular (cleavage and twinning) and physical (expansion) changes which promote disintegration through great heat.'^ 1) W. A. Tarr, ' Study of Heating-ïest.' Eco. Geol, 1915, p. 348. 2) During the field work in the volcanic district of Kaimon at the entrance of Kagoshima Bay in 1915, M. Nagabiichi made an interesting find of dull white subangtilar blocks of granite' about 10 cm. in diameter. The granite is trirtle-backed, much cleft and friable, and easily crumbles by hauimering. It is found scattered about between the northern slope of the konide of Kaimon and the northerly lying Lake of Ikeda. a caldron-shaped crateral depression of volcanic origin, nearly equal in dimensions and mass to the volcano of Kaimon. îno granitic rocks are exposed on the surface within a distance of 19 km. (in Prov. Osumi). Therefore, it is not unreasonable to conjecture that our fritted granite blocks were torn away and thrown out from the subcrust by an intense volcanic explosion, as in the case of éjecta of Sakura-jima. Here we have again the proof, that explosions of pent-up gases play a more important rôle in volcanic eruptions than we have hitherto imagined. THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. 219 12) Coal.— A fragment of coal was also collected by Kana near the western vent at an altitude of 300 m. It is a bituminous black coal which, however, shows submetallic lustre on the cleavage plane with a slight tarnish. One may be skeptical on this oc- currence near vents where we could scarcely expect to find it. There is no probability of coal having been carried up this barren mountain-slope by man before the eruption. As similar fragments have been picked up recently by other collectors, they may have been ejected during the first phase of the ti-uption from the bottom of ventholes, possibly from a lignite seam of the tuffite bed. Poor kinds of coal were reported to occur in the Plateau Formation of Tertiary age (?) from the neighboring provinces of Hyiiga and Satsuma. 13) Summary on the Loose Ejecta.— In summarizing what is said about the loose éjecta, we unexpectedly find a large number of them, both ancient and modern, and many present new features which must have uncommon interest for petrologists, especially ceramicites and gabbroids, the acid and basic pole of augite- andesites respectively. Both are leucolithic, and represent Yogt's anchi-monomineralic, extreme- differentiates of the andesite-magma. JNIore particularly, the projectiles of ancient heterogeneous origin are represented firstly, by pseudobombs of the ash-gray Kita-daké rock (pp. 212-214), and secondly, by trass composed of piperno andésite and its pumice of A and C zones (p. 14), besides the Mesozoic slate, and thirdly, by the so-called sandstone, being a glass-soaked pumiceous trachyandesite (pp. 15, 210). The second and third are, so to say, the volcanic recocts of the different members of the Plateau Formation of southern Kyushu. The fourth and the last one is the fritted granite of (?) Mesozoic age. The éjecta of preexisting rocks are exclusively throv/n out during the earlier phases of activity. 220 ART. 3. — B. KOTÔ : THE GREAT ERUPTION OF SAKURA-JIMA. On the whole, it is remarkable and quite astonishing to find among éjecta all the lithologie members of the snbcrust on which stands the volcano Sakura-jima. So far as the writer's knowledge goes, none of the substrata are omitted, and all are faithfully re- presented in the recent projectiles, so that we can read in them the geologic history of the underground of the region concerned. Lastly, the only one exception is the find of coal (p. 219) which is, however, of dubious origin. In closing the present treatise, the writer becomes aware of the fact that he has perhaps laid too miicJi stress on melting, stoping and gas-fluxing in contrast to hydrostatic pressure as to the mode of the ascension of the magma (p. 144). On seeing the multiplicity of the éjecta of subcrustal origin, the eruption seems to have started at first with a new rupture backed by gaseous ex- plosion, partly outside of the old conduit, whereby all the lithologie members of the foundation of the volcano were blown up through the virgin vents. CONTENTS. PAGE INTRODUCTIOX 1 PART I. General Outline of the Geologic Structure of South- west Japan 2 § I. GeogRxVphic and Geologic Sit l^atiox of Southwest Japax. 2 Ä. The Peri-Tangliai tectonic line and the (/encsis of the Rt/û- hyû arc 5 «) East coast of Korea 6 /j) Faulting of the axis oï Soutl) Japan 6 c) IMonoclinal Ejû-kyû arc 7 (1) Taiwan as a prolongation of the livû-kyû arc . . 9 e) Essential characteristics of the llvû-kyû volcanic chain. 10 B. The principal formations and main geologic structure of South- vest Japan 10 Tlie zones of «) Gneiss, 11 — h) Sambagawa-Mikabn Series, 11 — e) Cretaceous, 11 — d) Paleozoic graywackc, 11— c) Mesozoic sandv slate, 11—/) Tertiary, 12— g) Tiie lapilli plateau, i.e.. Plateau Formation . . . . 12 § II. Ox THE FORMATIOX OF THE TrEXCH BaY OF KaGOSHIMA. 13 A. General geology of the environs of Kagoshima Bay . . . . 13 o) Ash-stone (an Aso lava) . . . . 13 h) Origin of lapilli 14 c) Ash-stone and lapilli bed in tJie environs of the city of Kagoshima 14 d) The Seto region on the (.>pposite coast 16 e) Early Diluvial age of the ' lapilli plateau ' 17 /) How Kagosliima Bay was formed 17 g) The geology and dislocation of the Kaimon area. . . . 17 /() Concluding remarks on the formation of Kagoshima l^av 18 222 COXTEXTS. i) x\.ii epitome of the geologic history of the neighborhood <:>f Sakura-jima B. The outlying Osuml group Conclusion C. The foundation of Sakura-jima «) The six block islands : 1) Oko-shima, 23 — 2) Hakama- goshi, 25 — 3) The lava-drowned Karasu-jima, 26 — 4) Xanze, 27^ — ^5) Ivojima of Kokubu and its erroneous historic interpretation, 27 — 6) Yeno-shima h) Tlie subcrust of Sakura-jima PART II. The Volcano Sakura-jima § I. MORPHOGRAPHY ^4. General B. The building up of the volcano and topography (Appendix. — The vegetable zones of Mt. Sakura-jima) a) Kita-daké or the main, IS^orth Cone h) Naka-daké or the Middle Cone r) Minami-daké or the Soutli Cone d) Poison Gulcli, a mofette . . c) Polyaxial central type of c(jnduit /) Topographic feature . . g) Parasitic cones C. Hydrography a) Springs h) Mud-flood or Yama-sJiitco c) Strcamless § II. The Eruptions of Sakur.v-jima prior to 1914 .4. The eruption-periods I.— IV. (708-1766) tt) The eruption-period I. (708 a. D.-718).. 6) The eruption-period II. (764-766(?) ) . . c) The eruption-period III. TJic Bu.a[MEI eruption (1468- 1478) 19 20 22 22 30 30 30 31 31 35 35 3G 37 37 38 38 39 39 40 41 41 41 41 41 42 CONTEXTS. 223 (I) The eruption-period IV. (1642-176G) . . B. The cri(ption-2}€)'iod V. (1779-1799) a) The eruption of the Ax-ei era (1779-1780) b) Changes in the efflux of liot springs c) Ash-fall (t) The five new islands, 48 — Inoko-jima, 48 — Iwo-jinia 50 — Shin-jinia, 50 — Yebisn-jinia, 52 — Hama-shima ■e) Origin of the new islands /) AVaning phase of the An-ei activity a The eruption-period VL (1835-1879) § III. The Eruption of Sakura-.iima in 1914 (Eruption-period VII.) A. Premonitory Symptoms a) Hot springs 6) Tlie Yoshimatsu earthquakes c) Activity of Kirishima on Jan. 8tli d) The Kagoshima earthquakes on Jan. lOtli and lltli e) The avalanche of the Scissor RocJc /) Tiie flying of red-hot stones in western Sakura-jima -B. Surface manifestation of activity at Sakura-jima . . a) The first phase (Jan. 12-13th), 61 — TJie eventful day, Jan. 12th, 61^ — Hot and cold springs, 02 — Outbreaks inaugurated PAGE 43 44 44 47 47 52 53 55 56 56 56 57 57 58 58 59 00 61 The activity on the west side, 63^ — Height of the ash- cloud, 05 — The nuée ardente, 05 — Holes in the ground, 00 — Tlie first sign of lava-flow Tiie activity on tJie east (Nabé-yama) side, 07 — The 12th, morning, 07 — TJie rain of pumice in Seto, 08 — The great earthquake of tlie 12th, 0. 29 p.m., 09 — The tsunami b) Jan. 13th, morning, 71 — The nuée ardente, 71 — The first ash-fall in Kagoshima, 72 — TJie lava already 63 07 09 224 COXTEXT.S. ^ • PAGE covered the western slope at noon, 72 — The 13th, night, the climax of eruption 72 c) The second phase (Jan. 14- 20th), 73— The 14th, morning, 73 — îs^ew ventholes appeared successively on the icest, 74 — ^A^elocity of lava, 45 km. per hour. . . . 74 d) Jan. IStli, morning, 75 — ^ Velocity of lava-stream, 76 — The 15th, niglit, 77 — The eastern vents and the velocity of lava 77 c) Jan. 16tli, 78 — The western lava stream reached the sea, 78 — The Nabé-yama (eastern) vents and boiled fishes, 78 — Spatter cones 78 /) Jan. 17th, 79 — Tiie second ash-fall in Kagoshima turn- ing the city to night, 79 — Slight glacial period, 79^ — Ash-pisolite, 80 — ^Lava I'eached Karasu-jima, 80 — Lava filled the Seto channel reducing it to 18 m., 81^ — Ash-mud flood denied (footnote) 81 g) Jan. 18th, 82— The waiter's pioneer trip to the east coast, 82 — Lapilli buried houses on the Osumi coast, 82— Men buried alive, 82 — Pumice-asli flood, 83— Ter- minus of ' live ' lava at Fumoto 84 /i) Jan. 19th, 85 — Karasu-jima disappeared, 85 — Ash-fall, 85 — Peace was restored in Kagoshima 85 ?') Jan. 20th, 86 — Migrating pumicc-shect in the west, 86 — Southwest Sakura-jima suftcrcd little by the catast- rophe 86 j) The thied phase (Jan. 21 -27th), 87— Jan. 21st, 87— Rain of pumice on the east 87 k) Jan. 22nd, 87— Ash-fall in the city 87 ?) Jan. 23rd, 87 — -Nabe-yama vent I^o. 4 grown in size and great lava-flows, 87 — Suffocating gases on the A Osumi coast 88 m) Jan. 24th, 88— The 25th, 88— Odious gases again, 88 —The 26th, 89— Explosion on the east 89 CONTENTS. 225 PAGE n) Jan. 27tli, 89 — Apparent movement of tlie western lava-flow eamo to a standstill 89 o) ÏHE FOURTH PHASE (Jan. 28— Feb. 1st), 89— Jan. 28tli, 89— The 29th, 90— The 30th to 3 1st, 90— February 1st, 90— A new islet forined, 90 — Eventually Sakura- jima became a peninsula of Osumi, 90 — Feb. 2nd . . 90 2)) The fifth phase (Feb. 3 — Mar. 1st), 91 — Feb. 3rd to 4th, 91 — Tiie main vulcauism of Sakura-jima terminated, 91— Feb. 5th to 6th, 91— Feb. 7th to 8th, 92— The cloud- burst and mud-flood, 92— Feb. 9th to 12th, 92— Feb. 13th to 14th, 92 — Iwo-jima oif Satsuma quaked, 92 — Feb. 15th to 16th, 93 — The second ash-mud deluge in Osumi, 93— Feb. 17th to 18th, 93— Feb. 19th-22nd, 94— Feb. 23rd to 24th, 94— Feb. 25th to 26th, 94— Feb. 27th to March 1st 94 C. Economic and social statistics 95 D. Saimnary and co7iclusions 96 i. Predisposing causes 96 Recurrence of vulcanicity in general, 97 — Recurrence in Japan 98 ii. Premonitory symptoms 99 Earthquake-swarms of Yoshimatsu, 99 — Earthquakes and volcanic outbursts after maximal period of local atmospheric pressure 100 iii. The phases of eruptions 101 First phase, Jan. 12th to 13th, 101 — Second phase, Jan. 14th to 20th, 101— Third phase, Jan. 21st to 27th, 101— Fourth and fifth phase 102 iv. The mass and dimensions of lava 103 The western lava-field, 105 — The eastern lava-field, 105 — ^The whole mass of lava, 108 — Lava-delta and steam-spouting horn 108 — ^The ring of wet fumaroles. 110 v. The temperature of lavas 110 22G CONTEXTS. PAGE Melting raiioro of the Mihara-yania lav^a by Fuji and jNIizognclii, 111 — Melting experiments of the same by Tawara, 112 — Summary on melting experiments, 113 — Block-lava and bread-crust bombs, 114 — Secondary exhalations, 114 — Fumaroles, 115 — Lava-flows in sea, 115 — The tromba, 115 — -Temperature of sea- water, 116— Boiled fishes 116 vi. Lapilli, ashes and sands 117 The ashes, 117 — Crystal-lapilli, 118 — Magma-glass ashes, 119 — Tiic sands, 119 — ^Cauli flower-shape of dust-clouds, 119 — Resurgent and juvenile sands and ashes, 120— The lapilli, 120— Slag-lapilli, 120— Characters of eastern and western lapilli, 120 — Che- mical composition of ash, 121 — Acidity of ashes, 122 — Deposition of eje.-ta, 123 — Disjiersal of ejeata, 124 — Dust in the upper stratosphere, 126 — -Amount of éjecta 127 vii. Subsidence of tJie Sakura-jima environs 127 Relcvelings of the Ordnance Departuient, 128 — Curves of equal subsidence, 130 — The Mino-Owari earth- quakes of 1891, 131 — Cause of the subsidence, 133 — Smnmary, 135 — Local isostatic vulcanism 137 viii. Vcntholes 137 A view that the series of vents located on mere surficial fissure of the crust, 139— Western ventholes, 139 — Eastern ventholes, 142 — Meclianism of the re- cent Sakura-jima eruptions, 143 — Satellitic injeotion- clmmber 145 ix. The spatter eruption and the formation of a dome . . 146 The Yuno-hira dome, 147 — ^Mode of eruption of No. 5 vent of Nabé-yama 148 x. Sublimation-products 149 Lava-sublimates, 149 — Salmiac, the principal subli- CONTEXTS. 227 PAGE mate, !•")() — SuniDiaiy l-jl PAKT lir. Petrography ,152 § I. Tin-: TvAVAS op the PuE-insTORic Coxes 152 .1. The lai-a^i of KUa-daU 152 a) To]wgr:ipliic features, 152 — Northern slope, 152 — AYestern slope, 150 — The volcanello of Furuhata, 15o — The ao'glouierate bed that imderlies the North Cone, 153 — .Vtago-yama of dubious origin, 154 — The volcanic plug (^f Hikino-hira, 154 — Southern side, 155 — The ^Vkashi-Gongeii liill, 155 — Gongen-yania and Sono-yania on the eastern side 155 h) l'etrographi(^ cliaracters 155 B. T/ic Kabano lava (^iiorthcni side) 158 a) i\Iode of occurrence, 158 — The age doubtful -.159 h) Pétrographie characters 159 G The Jfinanti-dake and Nahi-dakc lams 160 a) ]\ro(le of occurrence, 160 — ^The Naka-daké tlow . . . . 161 />) Pétrographie characters, 161 — (i.) The Minanii-dake lava, 161 — (ii.) The Kannon-zald lava, 162 — (iii.) Tlie Naka-daké and the Nagasaki headland area 162 § II. The Lavas of the Historic Period 163 A. TJte lavas of the Bummel eruption, 1471-1476, including the Xahé-yama and Yebino-tsuha adventives 163 a. The Urano-mayé lava-field of 147 1 (eastern side.) . . 163 a) General 163 h) ]\Iode of octcuri'cnce 164 (') Pétrographie characters 165 ß. The Nabé-yama and Yebino-tsuha homates . . . . 165 a) The Nabé-yania homate 165 b) The Yebino-tsuka knob 166 B. The Bynbu-hira lava-field of 1475-1476 166 228 CONTENTS. TAGE a) ]Mocle of occurrence, 166 — The four pits, Nos. 1-4, 167 to 168— The vent No. 6 (the Taka-cliiki vent) . . 169 b) Pétrographie characters of joyroxene-andesite, 269 — TJie land and shore lavas 169 C. The Ohira-yama lava field of 1749 170 «) Mode of occurrence 170 h) Pétrographie characters of pyroxene-andesite 171 D. The An-ei lava of 1770-1781 172 o) Mode of occurrence, 172 — xin-ei-San, 172 — The Komen lava-field 172 h) Pétrographie cliaracters of hypersthcne-andesite, 174 — o) The Au-ei-San lava, 174 — /5) The Konien lava, 175 — Y) The rocks of the iVn-ei islands 176 III. General Petrographic Characters of the Lavas of 1914 176 A. 3Iasslve lavas 17<) 0) Land lavas in general, 176 — Hypcrsthene-andesite, 176 — The colors of lavas, 178 — The a and ß tyjx3S. . . . 180 b) Chemical characters 180 c) Submarine lava-flows, 181 — The lavas of the western field, 181— The lavas of the eastern field 183 fJ) Characteristics of the icestern and eastern lavas, 184 — Phenocrysts, 185 — The gronndmass 186 J^. The loose éjecta 187 1) Volcanic scums, 187 — Spongy thread-lace glass scoria, 187 — Ceramicites, a new group name, 187 — a) White scums, 189 — ß) Gray scums 187 2) The gabbroids as éjecta, 190 — -The a type, Microtinite, 190 — Tiie ß type, diopside-gabbro, 191 — The Y type, lithoidal microtinite, 192 — The d type, micronoritc, 193 — The s type, micro-allivalite, 194 — General explanation 195 CONTENTS. 229 PAGE 3) The cehamicites, natural porcelain ejocta, 196 — TIk; ürst type, a saeeharoidal niodifieation, 198 — ^ïlie second type, a vesicular lithoidal uiodification, 199 — ïlic third (normal) type, a compact lithoidal, porcelain-like modification, 200 — Chemical composition, 201 — The fourth type, 202 — ^The fifth type, a gray jaspilite-like modification, 202 — Origin of ceramicitcs . . . . 203-308 4) Solid ejected blocks ov .tuvexile lava 209 5) Bread-crust bombs 209 G) Recent lapilli (western and eastern) 211 7) Porphyritic obsidian 212 8) PSEUDOBOMBS OF ANCIENT LAVAS («, ß, )' and £ typCs) 212 9) Ejecta of trass or asli-stone, (a) balls, (/>) fragments. 214 10) The so-called sandstones (glass soaked lapilli.) . . 216 11) Ejected blocks of biotite-granite 218 12) Coal 218 13) Sü:sdiary on the Loose Ejecta 219 INDEX. Algean Sea oî Jai^an G Aerolite 191 Agglomerate bed of Mizu-otoshi . . 153 Air-qxiake (çua.st-earthquake) .... 66 Alaska (Katmai) 97 Amblystegite 113 Ambrym in New Hebrides 97 Ampferer, 0 133 Andersen, O., 178 Andrea. K 133 An-ei-jima (Moyé-jima, Shin-shima) . . 50 An-ei-San (Zan) 47, 172. 173 AnhydroTTs eruption ] 26 Anorthite of Myaké-jima and Tanamai 126 Anortbosite-eqitivalent . . ., ,. ., 191 Aoga-shima 98 Arimura 46 Asama 98 Ash 117 Ash, chemical composition of 122 Ash, lava — 119 Ash, magma-glass — 119 Ash-cloud, height of, 65 Ash-fall 47 Ash-fall, first in Kagoshima .... 72 Ash-fall in Tokyo 118 Ash-fall, second, 79 Ash-mud deluge in Osumi 93 Ash-stone, a hypersthene-trachyand- esite 15 Ash-stone (hal-isfU) 13, 14, 214 Asiatic bank 7 Aso, the volcano of, 4 Atago-yama lava 157 Avalanche of Scissor Eock, hasami . . 59 Azuma, K., 57, 146 Barren, J. 130 Benten Eock of Oki-Kojima .... 28 Bergeat, A., 198 Biotite-granite 187 Black rind of lava 214 Bocca, pi. bocclie 182 Bonin islands, afterglow and ash- fall 127 Bowen. N. L., 178 Brauns. R 198.205 Bread-crust bomb 114, 187, vOO Brown, C, 2 Bruhns 205 Bulla vernicona vur. ovula 16 ByôbXT-hira (Shibano) 169 Byôbu-hira lava, a pyroxene-andesite. . 169 Byôbu-hira lava of 1475-1476 ..160, 166, 167 Cabo di Gata 198 Campiglia Maritima 198 Card'mm muticum Ewe 16 Cardium retusum. L 16 Cdssis sp 16 Cendre granidée (pisolite) 80 Ceramicite 187 Ceramicite, chemical composition of, 201 Ceramicite, its formation, 205 Ceramicite, lithoidal 200 Ceramicite, natiural earthenware . . . . 197 Ceramicite, porcelain-like éjecta . . . . 203 Characteristics of the eastern lavas . . 184 Characteristics of the western lavas . . 184 Chlorahiminite 151 Chirique 97 Clino-enstatite 182 Coal 187, 218 Cognate xenolith 204 Colima 79 Cordierito 193,202,204 Cordierite, neovolcanic 204 Cordierite-liparite 198 Corona (rim) 19'.) Cretaceous 11 Crystal lapiUi (crystal sand) . . . . 118 Cupola 147 Daly, A., 53, 110, 144, 147, 213 Daikuhara, G 122 232 INDEX. Damage to property 95 Davison, C, 6^ Detonations 103 Diluvial 16, 29 Diluvial (early) 17, 19 Diopside-gabbro 191, 192 Domes, formation of, 146 Doro-sbima (Kani-sbima) !>2 Bosinia troscheli, Liscbke IG Dust-cloud (not asb-cloud) 59 Dust in tbe tipper atmospbere . . . . 126 Eartbquakes in Ijû-in 58,100 Eartbquakes in Iwô-jima 92 Earthquakes in Yosbim at su 57, lOU Eartbquakes, tbe greatest, 67, 69 Economics 95 Ejecta, deposition of, 123 Ejecta, dispersal of, 124 Ejecta of biotite-granite 216 Ejecta of biotite-granite of Kaimon . . 218 Enclave of diopside-gabbro 195 Enclave of diorite-scbist and gjibbro . . 196 Eruption, anhydrous 126 Eruption of Bummei (1468-1471) .. 42 Eruption of Cotopaxi of 1880 .... 127 Eruption, Icelandic, of 1875, . . . . 126 Eruption of Krakatoa 127 Eruption of Sawui 116 Eruption of tbe An-ei era 41, 172 Eruption of tbe Ansei or Mannen era 56 Eruption of the Empô era 43 Eruption of tbe Genua era 43 Eruptiou of tbe Ho-ei era 43 Eruption of tbe Hô-réki era .... 43 Eruption of tbe Kôji era 43 Eruption of tbe Kwam-po era .... 43 Eruption of tbe Kwan-sei era .... 56 Eruption of tbe Kwan-yei era .... 43 Eruption of tbe Kwan-yen era .... 43 Eruption of tbe Meiji era 56 Eruption of tbe Mei-wa era .... 43 Eruption of tbe Keiki era 4L Eruption of tbe Tembiô-Hôji era . . 41 Eruption of tbe Tembun era .... 41 Eniption of the Temmei era .... 55 Eruption of tbe Yôrô era 41 Eruption, phases of 1Ü7 Eruption, real, 72 Eniption, subnormal 145 Eruption-period I. (708-718) .... 41 Eruption-period II. (764) 41 Eniption-period III 42 Eruption-period IV 42 Eruption-period V 44 Eruptive type 149 Erythrosiderite 150 Eskola, Pentti, 208 Eucritic mass 194 Explosive type 149 Faience 202 Faulting of South Japan axis .... 6 Fifth phase 91 First jjbase of erui^tion 61 Fishes, boüed, 78, 116 Five New Islands 48 Flank eruption 38 Fo-kien, inchned tableland of, ... . 9 Foundered arena 110 Foundering eruption 147 Fouqué, F., 140 Foiirth phase 89 Framed folding {Rhamenfaltunfj) . . 8 Friedlaender, I., .. 21, 32, 44, 56, 116, 187 Fuji, N., Ill Fujisbiro 104 Fujiwhara, S., 66 Furuhata lava 147 Furubata volcanello 153 Fiiriisato 47 Futakami-yama, I'rov. Yamato . . . . 198 Futamata gulch 153 Gabbroids 187, 190 CamC) village 136 (jases, odius, 88 Gases, suffocating 88 Gas-blowing born 109 Gas-fluxing 220 Gifu 129 Gilbert, G. K. 58, 129, 130, 135 Glacial i^erioil SO GLihs cotton (Pele's hair} 80 Glomeroporphyritic 156 Glunger, G., 198 Gneiss 11 Gongén-yama 155 Gôro-yama, Prov. Shinano 197 Granite, fritted 30 Gravity, regional anomaly of, . . . . 130 INDEX. 233 Gray scums 189 Groundmass, chemical composition of, 196 Gypsiim 150 Gypsum, primary, 151 Hakama-goshi (Sbiro-yama) 20, 23, 29, 125 Hakama-goshi, shell bed of, .... 26 Hama-shima (Nii-shima) 52 Harker, A 204 Harker s cognate xenolith 208 Ilawaii (Puna) 97 Hébiga-kubo, ' snake's hollow,' . . . . 169 Héda-Kojima 28 Heights of ash-cloud 65 Henslop and Smythe 196 Higona (üre-hole) 53 Hikino-hira (Hyolai-oka) 38, 155 Hikino-taki bocca 171 Hobbs, W. H., 31 Holes in the ground 66 Holmgren, N., 22 Hornblende 157 Horses killed 95 Horst, relaxed, 7 Hot spring 47, 57, 87 Hot spring of Arimura 47 Houses buried 95 Hurricane, volcanic, 71 Hussak, E., 197 Hydrographie OfSce 105 Hydrography 39 Hypersthene-trachyandesite .... 21 Iddings, J. P., 110 limori 122 Ikéda, crater-lake of, 18, 218 Imamura, A., 45 Inland Sea {Seto-uchi) 6 Inoko-jima 48 Inoko-jima Eock 173 International Geodetic Committee . . 129 Ippai-jima (of Kokubu Kojima) .... 28 Iwakawa, T., 26 Iwô-jima (Sé-jima, Kuro-shima) . . 21, 50, 173 Izichi, S., 46 Jagger, T. A. F., 57, 98, 188 Jensen 61 Kabano lava, age luiknown . . . . 159 Ivixbano lava, vitrohypersthene- andesite 159 Kagoshima Bay 4, 8, 13, 18 Kagoshima Bay (100 fathoms) . . . . 16 Kagoshima Bay, a hollow 17 Kagoshima, trench b.ay of 13 Kaimon 4, 5 Kaimon, lowland of, 18 Kalinite 151 Kanai, M., 1, 26, 30, 117, 151, 191, 213, 216 Kani-shima (Doro-shima) 52 Kanzé 20. 23, 27, 29 Karasu-jima 20, 23, 26, 29 Katmai (Alaska) 97 Katô, ïak., 197 Kawabé Seven Islands 4 Kazuno 137 Kermadec group 97 Kilauea 53 Kilauea and Loa 144 Kita-daké (Mihachi) 34, 35 Ivita-daké hypersthene-andesite . . .. 155 Koike 25 Ko-jima (of Kokubu) . . . . 23, 27, 29, 41 Komaga-daké in Holdvaidô 193, 197 Komaga-daké, west of Morioka . . . . 197 Kômen 46 Kômen area 175 Korea, east coast of, 6 Korintji (Sumatra) 197 Koto, B., 6, 111 Kremersite 150 Kuchino-Erabu 21, 22 Kuro-shima of the An-ei islands . . 21, 22 Kuro-shima (Iwô-jima, Sé-jima) .... 50 Kyûshn, hörst of, 18 La Soufrière 206 Lacroix, A., 191, 193, 195, 197, 202, 913, 206 Lacroix, A., enclave homœgènes. . .. 208 Lapilli (karu-îshl or galshi) . . . . 117, 120 Lapilli buried houses 83 Lapilli, magma-soaked, 215 LapilH, origin of, . . 14 Lapilli plateau 14 Lapilli, recent, 187 Lasaulx, A. v., 205 Lava, block— 114 Lava, Furuhata — 147 Lava, temperature of, 110 Lava-ash (resurgent) 120 Lava-delta 108 Lava-flows, area of, 109 234 INDEX. Lava-flows, dimensions of, 104, Lava-flows, first sign of, Lava-flows, submarine, Lava-flows, the great flow in Nabé- yama Lava-flows, thickness of, 105, Lavas of 1914 Lavas of Kabano Lavas of Kita-daké 152, Lavas of Mihara-yama Lavas of pre-historic cones Lava-sublimate Lava-tunnel 109, Lesser Antilles Levelings Linschoten Is. (Cecilia Is.) ' Live ' lava Loose éjecta Loose éjecta. Summary on, Luzon, Taal Machatschek Magé-shima Magma ash (juvenile) Magma-glass ash Main viücanism terminated Martinique 1 Maskelynite Mass of lava jNIatav (in Samoa) INIatsura-gawara (gXilch) 153, Mauna Loa Mechanical analysis Mechanism of the Sakura-jima erup- tion Melting experiment Melting ii iter val Mesozoic Mesozoic slate Meteorological Observatory of Kago- shima Micro-allivalite Micronorite 156, 195, Älicronorilic segregation Microtinite 186, Middle Cone (Naka-daké) Milne, J., Misawa, M Minami-daké lava INIinami-daké lava, a pyroxene-andesite 107 67 181 107 176 152 155 111 152 149 110 1 131 4 76 187 219 15 8 22 120 119 91 , 97 194 104 97 158 53 117 143 112 111 11 31 4 195 211 171 190 35 44 150 160 161 Mino-Owari earthquake of 1891 . . 131 Mizogiichi Ill Molysite 150 Montague Pelée 200 Moya (mud-flow) 80 Moyédashi lava-field 173 Moyé-gona (• burning bocca ') . . . . 167 Moyé-jima (Shin-jima, An-ei-jima) . . 50 Moyé-saki 23, 42, 47 Mud-flow (moya) 80, 29 Mukomen 174 Miyazaki 72 Ifabé-yama 31 Nabé-yama, great lava-flow of , . . . . 88 Nabé-yama homate 165 Nabé-yama lava, chemical composition of 180 Nagabuchi, M 28 Nagaoka, H., 129 Nagasaki-hana headland 161 Nagoya 129 Naka-daké lava 160 Nakano-shima (Yebisu-jima) . . . . 52, 73 Nakano-shima lava 52 Nakashima, K., 31, 36, 44 New Hebrides (Ambrym) 97 New Islands, origin of, 53 Ninomiya, Doctor ■ . . 63 Nishio, K., 29 Nojiri lava 171 Nishio and Friedlaender 56 Kuéé ardente 65, 71 Nummulite {Pellastispira) of Motobu . . 7 Obora-yama, Prov. Isé 198 Obsidian, porphyritic, 187, 212 Oda. E,, 32 Ogawara ravine 153 Ogura and Matsiimoto 181, 201 Ôhashi, B., 201 Ôhira lava 153, 160 Ohira-yama lava-flow 43, 170 Oki-kojima (Oko-jima).. ..20, 23, 24, 26, 29 Oko-shima (OM-kojima) ..20, 23, 24, 26, 29 Olivine 185, 177 Olivine, resorbed, 178 Omori, F., . . 57, 101, 103, 106, 112, 127, 147, 183, 194 (Jmoyé-zaki 164 Ontaké of Shinano 193 INDEX. 235 Optrculimi (Tertiiiry) 13, 17 Ordnance Department 128 Osann, A 198 Ôsliinia of Idzii '.)8 Ôshima (Anami-Oshima) of Kyûkyîi 22 Ôseki, K 50 Ôsruni, tlie stniit (Yan Diemens Strait), of, 5 Outbreak inaugurated G3 Oxen Idlled 95 Öyu, M 198 A aleoz( ûc 31 Parasitic cone 38 Pelée, Mt 71 Pele's hair (glass cotton) 80 Peninsula, Saklirajima became, of Osumi 90 People killed 95 People resciied 95 People wounded 95 Peri-Tunghai 7 Peri-Tunghai tectonic une 5, G Perret, F. A . . . . 56, G5, 115, 140, 146, 187, 202, 208 Pétrographie jirovince of andésite . . Ill Petrography 152 Philii^piue Bay (South Japan Sea, Mariana Sea) 3 I'hos seuticosiis L. ]7 Pigs killed 'J5 Pillow lavas 181 Finus Tlambergü Pari 33 Piperno 13, 215 Pisohte (cendre granulée) 80 Plagioliimrite (rliyodacite) 14 Plateau Formation . . . . 12, 17. 27, 31 Poison gulch (Doku-dani) 37 Polyaxial central type 37 Population of Salmra-jima 95 Powers, S., 213 Predisposing causes ^6 Premonitory symptoms 99 Pseudobombs 187, 212 Pumice, floating. 68 Pumice, rain of, 68 Pumice-ash flood (yamashiro) .... 83 Piunice-desert on sea 81 Pumice- lapilli 61 Pumice-sheet, migrating 86 Pyrrhotite 198.199,204 Qi(r/.si-earthquake (air-quake) .... 66 Recent La])illi (little stone) . . . . 211 Reck, H., 20 Regional anomaly of gravity . . . . 130 Resorbed olivine 178 Rhamenfaltung (framed folding) . . S Rhyodacite (plagioliparite) 14 Riohthofen, F. v., G, 9, 12, 15, 18 Riverless 41 Rock salt 151 Ryûkyû arc 2, 4 Ryûkyû arc, monoclinal, 7 Ryiikyû, cordillera of, 3 Ryûkyû fore-deep 1, 8 Ryûkyû, genesis of 5 Ryûkyû Graben 7 Ryûkyû outcurve 9. 10 Ryûkyû, volcanic chain of , . . . . 3, 4, G, 7 Sakkabira (Ho-zaki) 16 SakTira-jima, now a peninsula .... 90 Sakura-jima, vegetable zone •), Fig. 2. — PJioti» liy Mr. Yamasln'ta, taken six years bef »re the recent (;ruption. See pages •V-i, oö. Fig. o. — A'icw from the castle-lull of Kagoshima. Photo hv iNtr. Yaniaguchi. See page -Vo. I Jour. Sei. Coll., Vol. XXXVIII., Art. 3, PI. II. V\ii. 1 ,.a Fi-. 2 Fig. 3 Yauiaguclii anil Sugiuioto photo. Koto : The Erpution of Sakura-jima. B. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE III. PLATE ni. Panoramic views of western Saknra-jiina before and after the recent eruption. Fig. 1. — Snow-mantled Sakura-jinia before the eruption, as seen from the castle-liill of Kagoshima, Plioto by ]Mr. Hidaka. See pages 23, 25, 27, •)0 Fig. 2.— Nearer view of Sakura-jinia after the eruption from tlie low flat islet of Ivanze. Photo by INIr. K. Yamaguchi in May, 1914, i.e., four months after the catastrophe. ïiie lava-flows, whicli encroached upon the sea in the foreground, was then still in an unconsolidated and fumarolic state, as uiay lie seen in the picture. See pages 30, 35, 30, 138, Jour. sa. Coll., Vol. XXXVIII.. Art. 3, PI. III. Koto : The Eruption of Sakura-jii B. KOTO : I THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE IV. PLATE IV. Fiji;. 1.- — ^Aiew of the cliara(;teristic land-feature of tlie Mapilli plateau ft)rma- tiou ' from Kokubu Railway Station, looking northwards toward the volcanic group of Ivirishima, which is faintly seen at a distance. See pages 15, 19, 29. Fig. !2. — View of tlie Kai mon lowland, as seen from the northern slo}x> of tlie volcano of Kaimon. On the left is the fault-scarp of the Benzaiten Gold ]\Iine, Avhose northern continuation marks tlie western border of the trench bay of Kagoshima. The right Jialf l)eyond the Kaimon loM'land in the foreground shows tlic large walled crater-lake of Ikeda, which is equal in size and volume to tlie \"o]cano Kaimon. The lake-surface is seen in tlie space left by a breach in the crater-wall. See page 17.~ in CL CO >< 03 B. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE V. PLATE V. Fig. 1. — View of tlio ash-niantled shallow crator-basin or rather sunken top of the Middle Cone in foreground, and the large engulfed solfataric crater of the South Cone beyond, as seen southwards from the top of the I^orth Cone or Kita-daké. Photo by ISlr. Yamaguehi on January 4th, 1915. Sec pages 35, 36. Yig-. 2. — Scene of the deserted Akobara villa2;e at the lava-front. Tiie oround is thickly sheeted with the so-called volcanic conglomerate — a mass of angnlar blocks of various sizes that rained contemporaneously witJi the ragiug fiery wind, nuée ardente, Avhic-h out rooted and barked trees. The hollows seen in the foreground were made by the impact of large block projectiles at the first phase of eruption. The ' live ' lava later flowed down and overrid the bewildered ground. Photo l)y the writer in INIay, 1914. See pages QQ, 71, 117. Fig. 3. — View taken on January 13th, 10'' 33™ A. ^r., 1914, showing the lava-stream api^roachiiig the shore near the Hakama-goshi Jiill on tlie western shore. Xote the rising vapor-cloud in the middle, s, from a spatter vent in the flowing lava. Photo bought in Kagoshima on January IStli, 1914. See pages 67, 72. Jour. Sei. Coll.. Vol. XXXVIII., Art. 3, PI. V. Fis. 1 Fig. 2 Fig. 3 Autlior and Yanuiguclii [ilintn. Kotü : The Eruption of Sakura-jima. B. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914 PLATK VI. PLATE VI. Yh^. ]. — The 'live' lava from the western vent was approaehing tlie Aka- niizn shore, leaving a margin of only 200 m., on the morning of January 15th, 1914. TJie text-figure 20 (p. 75) is the same view taken in tJie afternoon of the same day. Photo by tJie writer. See pages 15, 76. Fig. 2. — The saiiK! fluent lava reached the water's edge on January 16th, 1914, and caused the saline vapor to rise in dense curdy shape at the western lava-front on coming in contact NvitJi the >\-ater. PJioto by Mr. S. Nakasa. See pages 77, 78, 81, 82, 115. . l^'ig. •>. — The same lava advanced into the sea, and the islet of Karasu-jima now became a portion of the land on Jamiary 17th, 1914. PJioto bought at Kagoshima. See pages 78, 81, 82. Jour. sa. Coll., Vol. XX XV m., Art. 3, PI. VI. Fig. 1 Fisc. 2 i'iff. 3 Author and Nakasa phntn. Kotô : The Eruption of Sakura-jima. lî. KOTO : TBE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE VIL PLATE VIL pjo-. 1. — View of the lava-front in tlic sea, being enveloped by white enrdy clouds of saline fnnies, thus representing the Sawaiian stage in the erup- tion of jNIatav in 1905. Photo by Mr. S. Nakasa at 1 p. M., on January 18th, 1914, looking southwards from the top of the Hakama- goshi hill (in the o])posite direction to tJiat in PI. VI. Fig. 2). See pages 78, 115, 116. Fig. 2. — Snowing of the thiek ash-dust in Kagoshinia on January 17th, 1914. A few people were walking in rain-coats ^vith umbrellas, and a dozen soldiers were patrolling the streets in the same fashion to safeguard against fire and theft, as most of the houses were left vacant by the inmates who fled to the country. Photo bought in Kagoshinia. See pages 79, 80. Fig. 3. — View of eastern vent in eruption, a gale of dust stream ^vas coursing eastwards toward the writer who felt the air quite warm. A series of white saline fumes was risins: from black heads of the submarine lava which was flowing toward the opposite shore of the strait of Séto. Photo by the writer on January 18th, 1914, at Funioto in Prov. Osumi. See page 85. Jour. Sei. Coll., Vol. XXXVIII., Art. 3, PI. VII. FiiT. 2 FiiT. 3 Kakasa ami autlior iilwto. Koto : The Eruption of Sakiira-jima. B. KOTO: THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATK VIIL PLATE VIII. Fig. 1. — The western lava-flow drowned the islet of Ivarasu-jima on January 20th, 1914. Compare Fig. ;3, PI. Y. and Figs. 1 and 3, PI. YI. Photo bv the writer. See j^ages 85, 86. Fig. 2.* — AVestern view of the strait of Seto prior to the eruption, now entirely blocked up by the Xabe-yama lava-flows, so that Sakura-jima became a peninsula of Prov. Osumi. In the middle is seen the round- topjaed Akashi-(Tongén (103 on.), and on the left is the hot spring of Ari-mura, noA\- lava-drowned. Photo bought in Kagoshima. Sec pages 88, 105, 155. Fig. 3.* — ïlie same view of the now defunct strait of Seto, as seen from Seto village in Sakura-jima toward the Osumi shore. Photo bought in Kagoshima. See pages 88, 105, 155. * Cf. Plate IX. Figs. 2 and 3. Jour. sa. Coll., Vol. XXXVIII., Art. 3, PI. VIII. i''iir. 1 A Via:- 2 Fiff. 3 Kotô : The Eruption of Sakurti-jima. lî. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE IX. PLAIE IX. Fig. 1. — View of the -westerii lava-streains, both on land and in sea, from the Meteorological Observatory hill in Kagoshinia. The lava-flows then came to a standstill and yet remained unconsolidated, the ' foundered ' marine lava-field is encircled with a series of white fumes, as mav be seen in the picture. Photo by Mr. S. Nakasa, taken on April 4th, 1014. See page 80, 110, 116. Fig. 2. — View of the strait of Seto prior to the recent catastroj)he, looldng eastwards from Cape Tatsu-zaki. On the left is the round Akashi-Gongen hill (103 7??.) and the thriving little hot sjjring of Ari-mura. On the right is the butte-like isolated hill of Sakkabira on the Osumi side. It is a detached geologic block of the basal Lapilli Plateau Formation, left isolated during the formation of the Bay of Kagoshima. The same geologic body constitutes the foundation of Sakura-jima. Compare Figs. 2 and 3, PI. VIII. Piioto bought in Ivagoshima. Sec page 16, 88, 105, 155. Fig. 3. — Nearer view of the preceding (Fig. 2). Ari-mura in the foreground, the thriving hot spring and most prosperous village in Sakura-jima, now entirely disappeared owing to the overflooding by the recent lava. Photo bouglit in Kagoshima. See pages 16, 88, 105, 155. Jour. Sei. Coll., Vol. XXXVIII., Art. 3, PI. IX. Fier. 1 Fi.r. 2 Sakkaba Fiiï. 3 N;il< O 0 -is) ■cô CD o H. KOTO: THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE XVI. PLATE XVI. MieroseoiDic textures of the older effusives of the basal forma- tion and the K.ita-daké (North Cone) lavas. Fig, ]. — Biotite-vitrodacite or quartz-bearing biùtitc-liyper.stlioiK' trachyandesite, type «, from Oki-kojima. See pages V,\, 15, 24, 29, 'SO, löG. Fig. 2. — Splierulitic lioriiblendc-hypersthenc trachyandesite, type j-, from ()l)ama in Prov. Osumi. See pages 13, 15, 24, 29. Fig. 3. — Sphernlitic and perlitic liornblende-bearing obsidian of tlie liyjxir- stliene-trachyandesite, ty^x; ß, from Oki-kojima. See pages 13, 15, 24, 29. Fig. 4. — Holocrystalline trachybasalt (Aj in Text-fig. 3, p. 14) from the lava- droAMied islet of Karasn-jima. Fresh olivine as well as resorption-rests are seen in the fignre, the latter being represented by aggregates of magnetite "vvhieh looks very mueli like rhünite. See pages 25, 27. Figs. 1-4 are the pre-Sakura-jim.i effusives of the Laiiilli riatean Formation. Fig. 5. — Hypersthene-andesite with the phenoervsts of hyperstliene and plagio- clase, and microuoritic patches from the crater-margin f)f tlie top of Ivita- daké, the Xorth Cone. See pages 155, 157. Fig. 6. — Magnified texture of the same from another locality on tJie top-crater of Ivita-dake, showing mono-mineralic feldspar-microlites in tlie ground- mass. See pages 155, 157. Fig. 7. — Hypersthene-andesite of the Kita-daké rock-ty])e from the parasitic dome of Furuliata on the western slope of Ivita-dalce. See page 15S. Fig. 8. — Black vitrohypersthene-andesite with plexus of augite-microlites in the groundmass, representing the Kabano lava-flows on the northM'estern slope of Kita-daké. See pages 158, 159. -il Jour. Sei. Coll. Vol. XXXVIII., Art. 3, PI. XVI. .Z '^^ <^ r^ X135 .^£l99ik' . «s;-.- -l ^^- ■viv.^-^^ 1 t'^u:- .V "^-^ / A^r X ^;-.i ^^iM^^Mk!^ 1- i.)i t ;. 8 !34 mst Koto : The Eruption of Sakura-jima. li. KOTO: THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PJLATE XVII. PLATE XVII. Microscopic textures of the lavas of the Kabano lava-field, the South and Middle Cones, and the Bummei eruption of 1471-1476. Fig. 1. — Black vitrohypersthene-andesite of Kabano, the saîiic lava type as PI. XVI. l^'io-. 8, from wliicli this differs in its glassy base which is (liaracteristically dainascciied in texture. See jmges 158, 159. Fig, 2. — Pyroxene-andesite (hyp. > ang.) of the South Cone from Cape Kannon-zaki. Olivine was once observed (not in slide). Sec pages IGO, 161, 162. Fig. -">.■ — Pyroxene-andesite of the same type collected on the southeast slope of the South C^ine l)etweeii .']00-800 î/î. contonr-line. It is the normal type Avith abundant l)rown glass base. See pages 160, 161. Fig. 4. — Pyroxene-andesite of tlie same tyjx^ as the preceding, collected on the top of the ]\Iiddle Cone. See pages 160, 161, 162. Fig. 5. — Pyroxene-andesite of the INIiddle Cone, type ß, from the Nagasaki headland. See pages 160, 161, 162. Fig. 6. — Pyroxene-andesite la\:i of the Ihrimnei eruption. A sjxicinien from the lîyôbu-hira lava-field, vent Xo. 1, on tlie southwestern sliore. See pages 166, 170. Fig. 7. — Pyroxene-andesite lava of the same era from Cape Moye-zaki at tlie lava-cud (»f the Jîyôhu-liira flow. Sec pages 166, 170. Fig. 8. — Pyroxene-andesite at tlu; water's edge from the same locality as the preceding. See pages 166, 170. ^ '■^' ;^ photo. ^^X Jour. Sei. Coll. Vol. XXXVIII., Art. 3, PI. XVII. TiM^.'.^ ■ iiv^'iL' ■: i. t-T.% ■tofe,- M ,1 X135 ■♦■■<''**r ^^^ fïvx m >^ *v v« ^^ .^^Ä ^t •^■^: M'^ I I^^-- .^--^ Xl35 v> Koto ; The Eruption of Sakura-jima. 13. Koro : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE XVIII. PLATE XVin. Microscopic textures of lavas of the BumiTiei eruption of 1471- 1476, the Kwan-yen eruption (Ôhira-lava) of 1749 and the An-ei eruption of 1779-1781. Fig. 1. — Pyroxene-andesite (the gray type) from imdci- water at Cape Moyé- zaki, the lociility being the same as tJio Biimmel lava in PI. XVII. Fig. 8. See pages 16(3, 170. Fig. 2. — Pyroxene-andesite of the same era from a eraterlet in Urano-maye on the nortJieastern coast. See pages 163, 165. Fig. 3. — Micronoritie localization in the Bmnniei lava from under -water at Cape Omoye-zaki, tJie lava being the same as the pret'cding. See pages 163. 165. Fig. 4. — Pyroxene-andesite (ang. > hyp.) of the Ohira-yama lava of 1749 on the western slope of the South Cone. See pages 170, 171. Fig. 5.' — ïlie same magnified 120 diameters. See page 170. Fig. 6. — Olivine-hypersthene-andesite at the lava terminus at Cape Tatsu-zaki, flowed down from An-ei-zan(?) on the southern slope of the South Cone. See pages 172, 174. Fig. 7. — Hypersthene-andesite of the An-ei era at the water's edge at Fnru- sato, being a part of tlie preceding flow (p. 175). Olivine is not observ^ed in this rock, Ijut is characterized by abuudant augite microlite in the brown glass. See pages 172, 173, 175. Fig. 8. — Hypersthene-andesite of the An-ei eruption at Komen on the north- eastern coast. See pages 172, 175. Jour. Sei. Cell. Vol. XXKVIII., Art. 3, PI. XVIII. r photo. \'j 'i xi;i5 ' ' ^-.vr ^ y V-" '■'':' ^ ■■ , .^•/.// Fio". 4 8 'i3; Koto • The Eruption of Sakura-jima B. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE XIX. PLATE XIX. Microscopic textures of lavas of the An-ei eruption of 1779- 1781, and the recent eruption of 1914. JP'ig. ]. — Hypersthene-andesite of the xVu-ci eruption at the water's edtçe at Cajie llirabaiia on the nortlieastern shore. See ])age.s 172, 175. ^^ig- ~- — Hypersthene-andesite from Iwo-jima, one of the new islets that rose froiii the sea during the An-ei activity. See pages 172, 176. Fig. o. — Hypersthene-andesite of the recent eruption (11)14), taken from the lower portion of the la va- front of the western lava-field at Akobara. See pages 176, 177, 178, 179. Fig. 4. — Hypersthene-andesite from the same lava-field on its southern margin at Akaiiii/.u. See pages 176, 177, 178, 179. Fig. 5. — Hypersthene-vitroandesite with the damascened brown glassy ground- mass, ejected from a western vent, fell and made hollows in the ground at Alvôbara. The ejected block represents the hyaline type of lava, it encloses angular enclaves of whitish ceramicite, and makes veinlets in it. See pages 176, 179, 186. Fig. (). — Olivine-bearing hypersthene-andesite brought out from the bottom at 30 tkthoms at tJie lava-front on tJic Yokovama side. The olivine is not seen in the figure. See pages 176, 177, 181, 182. Fig. 7. — TJie same as the preceding, from the bottom at 10 fathoms on the Akamizu side. A resorption-rest of olivine is st;en in the figure. See pages 176, 177, 178, 181, 182, 183. (Figs. 3-7 are the photomicrographs of lavas from the western (front) lava-field.) Fig. 8.^ — Hypersthene-andesite at Yuno-hama in tlie eastern lava-field. See piges 176, 177, 178. Jour. Sei. Coll. Vol. XXXV I IL, Art. 3, PI. XIX. ^<Ä i^;^ #• .\-«¥i^;*»^'- :: l-i^r^^'^'^!^ ^^^v Ä" --'^•JîV. X' ï*_ V» Author photo. m: Koto : The Eruption of Sakura-jima. 13. KOTO: THE GREAT ERUPTION OF SAKÜRA-JIMA IN 1914. PLATE XX. PLATE XX. MicroscojDic textures of the lavas of the Nabé-yama (eastern) field, effluxeci in the recent eruption of 1914, and of the éjecta in the western field in the same eruption. Fig. 1. — Hypersthene-andesite ioriiiiuij; the lower portiou of the laud-lava at tJic lava-Jrowued watering-place of Ari-iuura. No olivine Mas detected, l)ut angite is fairlv abundant among ])lieuo('rvsts. See pages 176, 177. Fig. 2. — Olivine-hypersthene-andesite from a lava head within the sea near Ari-mura. On account of higli magniti cation, unfitrtunately no pheno- erysts of olivine, hypersthene nor j)lagioclase are seen in the mieroseopic field. See pages 176, LSI, 182, 183. Fig. o. — Olivine-hypersthene-andesite of sul^marine origin like the preceding, forming an islet oft' shore of Ari-jniira. See pages 176, 177, 181, 182, 183. Fig. 4. — Olivine-hypersthene-andesite from the lava-front at the Avater's edge at the defunct strait of Seto. See pages 176, 177, 181, 183. (Figs. 1-4 are the pliotomicrographs of lavas ot the eastern (Nabé-yama) lava-fiekl.) Fig. 5. — A secretionary patch in the gray hornblende-andesite of Fiiruhata, a voleanello adventive to the North Cone. It occurs scattered abont in Akobara as éjecta or pseudobombs, having spindle-shape due to subaërial abrasion. It sliows the strucfHir encherêfrêc See page 213,. Fig. 6.- — White scum from the western coast with pores greater than 0.0'^ inm. in diameter. See pages 118, 18λ, 202. Fig. 7. — Gray scum from the same region as the preceding with i>ores less tlian 0.03) v//v/i. in diameter. See pages 118^ 189, 190. Fig. 8. — Microtinite éjecta on the western lava-field coated black witli lava films. It is built up of clear anorthite polygons cemented with grayish ])at(lies of (?) ortlioelase full of air pores. See pages 190, ]î)l. Jour. Sei. Coll. Vol. XXXVIII., Art. 3, PI. XX. Xl5u ^-:-^# Uitlior plioto. Kotü : The Eruption of Sakura-jima. ]i. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE XXL PLATE XXL Microscopic textures of various éjecta, namely ; scums, micro- tinites, mieroallivalites and gabbroids, ejected during the eruption of 1914. Fig. 1. — Cordieritiferous scum «>i- ceramicite-pninice from tlie western shore, h■ll0^ving cordierite aggregate filled with globules of pyrrhotite, the whole heilig enclosed bv a fringe of bladed air-pores. It is a unique specimen o^ pumice vitli cordierite. Piioto bv Asist. Professor M. Oyu. See j^ages 190, 200, 202. Fio-. 2. — Microtinite, being tJie same as in PI. XX. Fig. 8, seen under crossed niçois. Sec ])ages 190, 191. Fig. 3. — Micro-allivalite, which occurs as a round body coated M'ith lava-film, and is found as blocks on the margin of eastern vents during the late phase of activity. See pages 190, 194. Fig. 4. — The same under crossed niçois. See pages 190, 194. Fig. 5. — Gabbroid éjecta from the eastern slope of the Soutli Cone. See pages 190, 191. Jour. Sei. Cell. Vol. XXXVIII., Art. 3, PI. XXI. VC4 «^ly ^^ X17 Koto : The Eruption of Saiaira-jiina. B. KOTO: THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE XX 11 PLATE XXII. Microscopic textures of éjecta of the gabbroids, and the cordie- rite-bearing secretionary products or eeramieites erupted in 1914, except Fig. 8, which is an. Asama ceramicite of 1913. Fi«'. 1. — Gabbroid nuiguifiod 79 diameters. The suiu! éjecta as PI. XXI. Fig-. 5. See pages 11)0, 191. l^'g. '2. — Ceramicite or eordierite-bearing éjecta, type I., found at Koike. The corroded anortiiite with a clear gla.'^.sv rim enclosed in large cordierite plates full of minute air-pores. See pages 196, 198, 199, 205. Fig. o. — Ceramicite, type IT., under crossed niçois, from Kurokami <»n tJie eastern sjiorc.. \ eordierite crystal in the centre is built up of heteroge- neous grains of the same with homogeneous extinction of light, and pyrrliotite serves as the infilling matter. See pages 196, 19S, 199, 200^ 205. Fig. 4.— Ceramicite, tyjx* 111., from the same locality as above. A large crystal of cordierite is for the greater part built np of similarly orientated rectangles of the same mineral. See pages 196, 198, 200, 205. Fig. 5. — Ceramicite, tyjje IT!., with porphyritic polysyuthetic anorthite, seen under crossed niçois, from the same locality as the two 2)reeediug. See pages 196, 200, 201. Fig. 6. — Ceramicite, type IV., from spongy and banded eje;;ta near the Yuno- hira vent on the western side. See pages 196, 202, 205. Fig. 7. — Ceramicite, type V., from the same locality as above. It is a com- plex of minute rectangles of cordierite mixed with prismoids of plagio- clase and hypersthene. See pages 196, 202. Fig. 8. — Ceramicite, from the volcano Asama, showing corroded anorthite in the ground which is composed of rectangles of cordierite cemented with pyrrhotite, showing cliaraeteristically the so-<.;alled .v'dcronitic texture. Cordierite and pyrrliotite here seem to have crystallized nearly at the same time. See pages 196, 204. Jour. Sei. Coll. Vol, XXXVIII., Art. 3, PI. XXII. iutlior ijhuto. Koto : The Eruption of Sakura jima. li. KOTO : THE GREAT ERUPTION OF SAKURA-JIMA IN 1914. PLATE XXIII PLATE XXIII. Microscopic textures of various ejectamenta thrown out during the eruption of 1914. Fit»'. 1. — ^Fritted friable granite-block, iiiulcr (a-os,sed iiiculs, ejected from the hidden foundation of the volcano. Quartz is cracked, ortlioclase pseudo- phitized and biotite impregnated with i)yrrliotite along cleavage plane. See pages 30, 216. Fig. 2. — The so-called sandstone, which is really a glass-soaked pumice layer of the Plateau For j nation of southern Kyushu. Biotite is fresh (not found in the photomicrograph), but the rhomb-sJiaped hornblende is opacitized. See page 216. rig. '">. — Trass or a loose aggregate of heterogeneous constitution, mainly built up of lapilli of biotite-hornblende andésite soaked witli opal-like isotropic substance. Sec .30, 212, 214. Fig. 4, — Porphyritic obsidian of hyjiersthene-andesite, found as blocks at an altitude of 400 m. above the coastal village of Ari-mura. The brown glass in the groundmass shows the characteristic kneaded and damascened texture. See pages 186, 212. Fig. 5. — Pumiceous lapilli with hypersthene that rained at the Experimental Station of the Ivagoshiina Agricultural College, at a distance of 11 km. from the Nabé-yama vents. See page 211. Fig. 6. — Crystals of hypersthene, plagioclase and clumps of magnetite from pressed lapilli that thi(;kly cover the coast of Prov. Osumi (Tabé). See pages 118, 211. Fig. 7. — Wash-skeleton of ash that fell at Kokubu Station during the first phase of activity, at a distance of 20 hn. Minerals found are idiomorphie hypersthene, allotriomorphic plagioclase and a subordinate quantity of magnetite clumps. See page 118. Fig. 8.- — Ash that fell in Tokyo, at a distance of 1,000 hii. from Sakura-jima, being mainly composed of minute sjilinters of colorless strained glass full of air-pores. It is the lightest portion of connnimited lu]>illi. See pages 72, 118. Jour. Sei. Coll. Vol. XXXV III., Art. 3, PI. XXIII. Author photo. ^i^Vi À.3) l«% ■^ ^IpÄ* -1.4^ ^^ \^ XM^.^ Koto : The Eruption of Sakura-jima. GEt)LOr.Tr NfAP OK VOLCAXO SAK'riîAJIMA BY B.KOTÔ, PH D. Mesozoic Slate Basal Plateau iïu Plateau Formation Middle (lone-lnva SouUi Conclavii i-ei Lava 1779 New Land Lava JOURNAL OF THE COLLEGE OF SCIENCE, TOKIO IMPERIAL UNIVERSITY. VOL. XXXVIII., ARTICLE 4. Contributiones novae ad Floram Bryophyton Japonicam. PARS SECUNDA. Elaboravit Shûtai Okamura, BigakuhalmsM. (Institutione Botanico Universitatis Imperialis Tökyöensis). Cum 42 figuris. I. HEPATICAE. I MARCHANTIAOEAE. Meboulia haeinisx>1iaerica (L.) Kaddi. \ Formosa : Prov. Tailioku-cliô, Urai [.^ 5R] (Leg. Jukichi Shiraga ! VII. 1911.) ; in monte Arisau (Leg. BüNZö Hayata ! 6. IV. 1914.). \ Species nova ad floram formosanam. ' MarcJiantia tosana Steph. Formosa : Prov. Tailioku-cho, Urai [.% Mi (Leg. JuKlOHl Shiraga ! VII, 1911.) ; in monte Arisan (Leg. BuNzô Hayata ! 6. IV. 1914.). Species nova ad floram formosanam. ART. 4. — SHÜTAI OKAMURA JUNGERMANIAOEAE AKROGYAE. Plagiochila fruticosa Mitt. Formosa : In monte Arisan (Leg. BuNZö Hayata ! 6. IV. 1914.) Species nova ad floram formosanam. Chiloscyphus viviilaris (Schrad.) Loeske. Abh. Bot. Ver. Prob. Brand, pp. 172-174 (1904). Syn. Chiloscyphus polyanthus (L.) Corda, var. rivularls (Schrad.) Nees. Eur. Lab. n. p. 374. (1836). Habitat in solio lacus c. 2ni. prof undo. Hondo : Prov. Kôduke, in Lacu Marunuma [% îS] (Leg. Haraikawa ! VH. 1914.). Species nova ad floram japonicam. Bazzania spinosa Sh. Okamura. Sp. nov. Plantae robustae, caespitosae, lutescenti-fuscescentes, hand nitidae, rigidiusculae. Caulis repens, usque ad 5 cm. longus, cum foliis c. 3 mm. latus, dichotome ramo- sus, hie illic flagellaris, fla- geliis usque ad 3 cm. longis, microphyllinis, per totam lon- gitudinem dense fasciculatim fere hyalino-radiculosis. Folia dense imbricata, oblique-paten- tia, e basi semi-cordata ovalia, apice rotundato-truncata, con- cava, c. 1.6 mm. longa, basi c. 1.0 mm. et apice c. 0.6 mm. lata, marginibus anticis subplanis integris, posticis incurvis e basi ad apicem densiuscule (10-16) spinosis, spinis c. 80-250/^ longis, e seriebus cellularum basi 2-3 ad apicem 1 compositis, cuticula A. X20 B. Fig. 1. Bazzania spinosa. A. folium. (x20). B. Amphigastrium. (x20). CONTRIBUTIONES NOVAE AD FLORA!« BRYOPHYTON JAPONICAM. 3 spiiiorum incrassatia ; cellulis basilaribus rectangulis c. 35-42 [i. longis et c. 15-20,« latis, in medio et apice folii plerumqne quadratis, c. 20-35/^ magnis, marginalibus quadratis c. 14-20 j« magnis, omnino laevissimis valde chlorophyllosis incrassatinsculis, trigonis parvis vel nullis. Amphifjastria magna, dense imbricata, transverse inserta, decurrentia, cordato-quadrata, valde recm^vo-concava, c. 0.8-0.9 mm. magna, marginibus recurvis, ubique crebre spinosis. Formosa : in monte Arisau (Leg. BuNZô Hayata ! 4. IV. 1914.). Nomen speciei ab spinis marginis folii. TricJiocolea tonientella (Hubs.) Lindb. Kyushu : in insula Yakushima (Leg. Takanori Iwaki ! 21. IX. 1914.), Radiila complanata (L.) Dum. Formosa ; Prov. Taihoku-cliô, Urai (Leg. Jukichi Shiraga ! VU. 1914.). Species nova ad floram formosanam. Madotheca Perottetiana Mont. Formosa : Prov. Taihokn-chû, Urai (Leg. JuKicm Shiraga! VII. 1914.). Species nova ad floram formosanam. Fvullania inonlliata Nees. Formosa : Prov. Tailioku-chO, Urai (Leg. Jukichi Shiraga! VIE. 1914.). Species nova ad floram formosanam. ART. 4. — SHÜTAI OKAMURA II. MUSCI. SPHAGNACEAE. Sphagntun acutifoliiirti Eheh. Sachalin : Sicca [Sfc #] (Leg. Jüsuke Susuki ! 1. V. 1913) ; Ocliiai (Leg. HaeüZü Ivomatsit ! Till. 1914.). Korea : Prov. Kankyô-Nandô, Sanyo [Uj ^] (Leg. TAKEKOSHm Nakai ! 22, VU. 1914.) ; in monte Hôtaizan [^ä fà Uj] (Leg. Takenoshik Nakai ! 6. YHI. 1914.) ; in monte Koliô [^ mf.] (Leg. Takenoshin Nakai ! 10. VII. 1914.). Species nova ad floram koreanam. S2>hagnum Girgensohnii Russ. Sachalin : Sicca (Leg. Jüsuke Susuki ! 1. Y. 1913.) ; Ocliiai (Leg. Haeuzö Ivomatsu ! Vni. 1914.). Korea : in monte Eohô (Leg. Takenoshin Nakai ! 10. YII. 1914.) ; in monte Hôtaizan (Leg. Takenoshin Nakai ! 6. Vni. 1914.) ; in monte Hakntôzan [â IH Ui] (Leg. Takenoshin Nakai ! 10. YHL 1914.). Species nova ad floram koreanam. Sphagniiin japoniciim Warnst. Sachahn : Sicca (Leg. Jüsuke Susuki ! 1. Y. 1913.). Species nova ad floram sachalinam. Siyhagnnm papillosuin Lindb. Sachalin : Sicca (Leg. Jüsuke Susuki ! 1. Y. 1913.). Species nova ad floram sachalinam. Sphagman sqiiarrostnn Pers. Korea : Prov. Kankyö-Nando, Shindôrei [it it a]. Cliöshin-gim (Leg. CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 5 Takenoshin Nakm ! 13, VII. 1914.) ; in monte Roliô (Leg. Takenoshin Nakai ! 10. Vn. 1914.). Species nova ad floram koreanam. DIORANACEAE. Trematodon di'ejxmellus Besch. Formosa : Taihoku (Leg. Hisahiko Sasaoka ! 7. III. 1912. et Yaichi Shimada! IV. 1913.). Tt^ematodon longicollis Michx. Formosa: in monte Arisan (Leg. BuNZö Hayata ! 6. IV. 1914.). Species nova ad floram formosanam. Ceratodou purpureus (L.) Brid. Korea : Prov. Heian-Hokndo, Hiraihô [^ ^ ^] (Leg. Takenoshin N.^KAi! 10. VL 1914.). Species nova ad floram koreanam. Dicranella hetevonialla (Dill., L.) Schevip. Formosa-; Prov. Taihoku-chô, Shabosan [v'l? iff ilj] (Leg. Yaichi Seviada ! 31. xn. 1913.). Species nova ad floram formosanam. * Dicranuni Bergeri Blandow. Syn. D. Schraderi Schwör. Sachalin : Sicca (Leg. Jüsuke Süsuki ! 1. V. 1913.). Species nova ad floram japonicam. Dicranuni niajus Turn. Sachalin : Sicca (Leg. Jüsuke Süsuki ! 1. V. 1913.). Korea : in monte 6 ART. 4. — SHÜTAI OKAMÜRA : Kollo (Leg. Takenoshin Nakai ! 11. YII. 1914.). Species nova ad floram koreanam. JDicramirn scoparlum (L.) Hedw. Korea : Prov. Heian-Hokudö, Hoknjômen, Kokai-gnn (Leg. Takeno- shin Nakai! 1. Tu. 1914,); in monte Kohô (Leg. Takenoshin Nakai! 11. Vn. 1914.). Species nova ad floram koreanam. FISSIDENTACEAE. J^Hssidens japonictis Doz. et Molk. Formosa : Taihoku (Leg. Yaighi Shimada ! 20. VI. 1914) ; Ural (Leg. JüKiCHi Shieaga! vn. 1911.). Species nova ad floram formosanam. Fissidens lateralioides Sh. Okamura. sp, uov. Habitat in terra irrorata. Plantae teneUae. laxe caespitosae, laete virides, haud nitidae, moUiuscnlae. Caulis adscendens, c. 7-10 mm. longus et c. 0.16 mm. diam., cum foliis c. 2.5-3.0 mm. latus, infima basi fusco-radiculosus, simplex, densiuscule foliosus, obtusus. Folia sicca leviter crispata, madida erecto-patentia, c. 10-15-juga, infima c. 4-juga minuta, caetera majora, hand decurrentia, lineari-lanceolata, apice acuta, c. 2.0-2.3 mm. longa et c. 0.5-0.56 mm. lata, integerrima autem summo apice minutissime serrulata, lamina vera usque ad medium folii producta, oblique truncata, lamina dorsali ad basin nervi enata, limbo ad apicem producto, hyaline, dorsali inferne latiore c. 30-35 /< lato e seriebus c. 5-7, superne sensim angustiore e seriebus duabus cellularum elongatarum composite, ventrali inferne e seriebus 2-3, superne e seriebus duabus composite ; nervo concolore vel saepe fuscescenti, superne CONTDIBUTIONES NOVAE AD FLOKAM BRYOPHYTON JAPONICAM. stricto breviter excedente ; cellulis valde chlorophyllosis, laevibus, pellucidis, hexagonis, in medio folii c. 9-14;" magnis, superioribus minoribus c. 7 /^ magnis, basilaribus ad nervum rectangulis vel fere rhomboideis c. 20-25/^ longis. Inßorescentia antoica. ; flores masculi geminif ormes ; antheridia c. 10. Ramulus perichaetialis fusco-radi- cnlosus. Bracteae perichaetii in- timae basi alte vaginantae line- ares, apice breviter attenuatae, c. 1.2-1.4 mm. longae et c. 0.12 mm. latae, marginibus basi late superne anguste ex unica série cellularum formate limbatis, e medio ad apicem remote crenulatis ; nervo valide, infra summum apicem evanido ; cellulis hexagonis c. 10/^ magnis, basilaribus linearibus, in- fimis rectangulis. Vaginula anguste oblongO-Cylindrica, c. 0.3 mm. alta, C- Bractea perichaetiaHa intima (x35). D. Sporangia cum seta et vaginida ( x 20). fusca. Seta lateralis c. 1.3 — 1.6 Fig, 2. Fissideiis lateralioides A. Planta fertiüs ( x 1) ; B. Folia ( x 20) ; mm. alta, strictiuscula vel inferne arcuata, lutescenti-rubra, laevis. llieca erecta, oblongo-cylindrica, symmetrica, brevicoUis, c. 0.5-0.72 mm. longa et c. 0.32 mm. crassa, laevis ; ceUulis exothecii rectan- gulis, c. 30-35/^ longis et c. 15/^ latis, parietibus incrassiusculis, ad orificium in seriebus c. 5 minutis hexagonis ; stomatibus nuUis. Anmilus nullis. Peristoma dentés subulati, basi connati, ad 3/4 profunde bifidi, c. 0.32 mm. longi et basi c. 35 /^ lati, rufescentes, basi dense trabeculati minutissime papillosi, intus dense et alte lamellosi, cruribus duobus, filiformibus, dense spiraliter incrassatis. Sporae c. 7-9 ,« magnae, virides vel ferruginei, laeves. Operculum e basi conica longe rostratum. c. 0.35-0.4 mm. lono;um et basi c. 8, ART. 4. — SHÛTAI OKAMURA : 0.25 mm. diam., rufescens, rostro stricto vel curvatulo. apice obtuso. Calyptra cncullata, c. 0.56 mm. longa, viridescens vel pallida apice fusca. laevis. Hondo : Prov. Etclm, Kamidaki-mura, Kamishinkawa-gim (Leg. HlSA- HiKO Sasaoka ! 12. XI. 1914.). Habitus, statura, feciesque omnino F. later all Bkoth., foliis autem e cellulis majoribus, limbo latiore, inflorescentia autoica, seta breviore, theca oblongo-cylindrica diversa. Nomen specie! ab fecie ut F. lateralis Beoth. Fissidens nagasakinus Besch. Formosa : Taihoku et Mnbikö [M M W (Leg. Hisahiko Sasaoka ! 30. X. 1912.). Fissidens zippeliaiius Br. Jav. Formosa : Prov. Shincliiku-chö, Chikunan-ippo [1t ^ — S] (Leg. Hisa- hiko Sasaoka ! 21. X. 1912.). POTTIACEAE. Weisia viridula (L.) Hedw. Formosa : Prov. Shincliiku-chö, Chikunan-ippo (Leg. Hisahiko Sasa- oka ! 21. HE. 1912.) ; Prov. Taihoku-chö, Mnbikö (Leg. Hisahiko Sasaoka ! 3. XL 1912.). ßarbula orientalis (Wies.) Broth. Formosa : Taihoku (Leg. Hisahiko Sasaoka ! 1. XI. 1912.). Tortula einarginata (Doz. et Molk.) Mitt. Formosa : Prov. Shinchiku-chö, Chikunan-ippo (Leg. Hisahiko Sasa- oka ! 13. VL 1912.). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. GRIMMIACEAE. GlyphottiitrUini angustifoliuni Sh. Okamura. sp. nov. Habitat in saxis. Plantae robustae, compacte pulvinatae, su- perne laete virides, inferne nigrescentes, in statu sicco nitidiusculae, nigrescenti-virides, rigidiusculae. CauUs erectus vel adscendens, usque ad 1.5 cm. altus, basi rubiginoso-radiculosus, simplex vel 1-3-ramosus, sectione teres, c. 0.24-0.28 mm. diam., fasciculo cen- trali c. 40// magno pluricellulari, reti inter- medio fuscescenti, cellulis hexagonis vel oblongo-hexagonis, c. 15-20 ;" magnis tenellis, cellulis periphericis in seriatis 1-3 minoribus c. 7 !J- magnis incrassiuscuKs fuscis ; ramis erectis, c. 5-8 mm. altis, strictis, simplicibus ; caulibus ramisque dense foliosis teretibus. Folia sicca circinato-incurva, madida erecto- patentia, carinato-concava, lanceolata vel subulate -lanceolata, apice obtusiuscule acuta, c. 4.0-4.5 mm. longa et c. 0.64-0.72 mm. lata, marginibus integerrimis, undulatis ; nervo valide, apice evanido, in sectione transversali piano- convexo, dorso valde prominenti, basi c. 0.12 mm. lato et c. 0.06 mm. crasse, cellulis ventralibus c. 11-12 majoribus, ducibus medianis c. 7, stereideis numerosis, cellulis dorsalibus c. 17 distinctis minoribus ; cellulis laminalibus incrassiusculis, valde chlorophyllosis, obscuriusculis, ad apicem rotundato-hexagonis, c. 12-14// magnis, ad basin quadratis vel breve rectangularibus, basilaribus infimis rectangularibus, c. 30-40 // longis et c. 12-17 // latis. Inflorescentia autoica ; flores masculi ad basin perichaeti A.XIO B.y2o 7*7■ magno, reti ceterum incrassiore, areolis nbiqne fere ejusdem magnitndinis, celliilis hexagonis c. 15^ magnis fuscis c. 3-4-seriatis ; ramis brevibns, vix ultra 2 mm. longis, erectis, simplicibus vel innovando-ramulosis, dense foliosis. Folia sicca plus minus inflexo-adpressa, madida patentia, carinato- concava, lineari-lanceolata, breviter acuminata, c. 1.1-1.4 mm. longa et c. 0.24-Ö.28 mm. lata, marginibus usque ad medium re- volutis, integerrimis ; nervo crassiusculo, apice evanido, in sectione transversali piano -convexo, dor so paulum prominenti, basi c. 42 /^ lato et c. 20 j« crasso, e stratis 2-3 cellularum uniformium com- posite ; cellulis laminalibus majoribus, minutissime papillosis, haud opacis, rotundato-hexagonis, c. 14-17 ij- magnis, parietibus crassius- culis, in medio folii utrobique e nervo ad margina c. 7-8-seriatis, CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 17 A.X30 X20 FifJ, 7 . Aulacomitrivun miiiutissimum. A. Folia ( X 30) ; B. Theca cum Bracteis perichaetii intimis (x20). Vaginula eylindrica, c. basilaribus ntrobique c. 8-10-seriatis, ad nervum rectangulis c. 30-35 (j- longis et c. 14 !'- latis, ad angnlos quadratis c. 14-15/i magnis saepe fuscescentibus. In- ßoresccntia antoica ; flores masculi axil- lares ; folia perigonialia intima minuta, ovata, apice acuta ; antheridia c. 5. Bracteae perichaetii intimae longissimae, apice breviter acuminatae, c. 1.3-1.5 mm. longae ; nervo tenui, ad 1/2 evanido ; cellulis anguste rectangularibus, c. 20-35 /^ longis et c. 7-8 >« latis, basilaribus longiori- bus, superioribus brevioribus c. 14-20/^- longis et c. 7-10/^ latis, omnibus laevibus. 0.32 mm. alta. Seta c. 1.1-1,3 mm. alta, erecta, tenuis, lutescenti- fuscescens vel fuscescens, laevis. Theca erecta, ovalis, c. 0.4-0.5 mm, longa et c. 0.32-0.4 mm. crassa, deoperculata trugide ovalis, fusca, laevis ; cellulis exothecii valde incrasatis, ad orificium in seriebus c. 5-6 transverse rectangularibus valde incrassatis fuscis. Peristoma dentés per paria approximati, lanceolati, c. 0.16 mm. longi et basi c. 50-56 /^ lati, densiuscule articulati, minutissime papiUosi vel fere laeves, fusci. Sporae c. 30 ^ magnae, fascae, papillosae. Opercurum e basi conica rostratum, c. 0.32 mm. longum, rostro erecto apice obtuso. Calyptra campanulata, thecam totam obtegens, c. 0.96 mm. longa, luteo-fusca, pluries (c. 10) plicata, nuda. Hondo : Prov. Sagami, in monte Hakone (Leg. Kiyotaka Hisauchi ! 3. I. 1915.). Species distinctissima, a congeneribus statura minore, foliis angustis, cellulis folii majoribus facillime dignoscenda. 18 ART. 4. — SHÜTAI OKAMURA : Nomen speciei ab statiira minutissima plantae. Macromitrhini ificurviini (Li\db.) Par. Formosa : Prov. Taihokn-chô, Sözansliö [^ Ui St] (Leg. Yaichi Shima- DA ! 1. I. 1914.) ; Taihoku (Leg. Yaichi Shimada ! 2. VI. 1914.) ; Mentenzan [S 5S Uj] (Leg. Yaichi Shimada ! 31. X. 1914.). FUNARIACEAE. Physcomitrium eurystoniiifn (Nees.) Sendt. Formosa : Taihoku (Leg. Hisahiko Sasaoka ! 29. U. 1912.). Species nova ad florara formosanam. JPhyscomitrium subeurystoniuai C.4.RD. Formosa: Taihoku (Leg. Yaichi Shimada! 20. "NTL, 1914.). Fiinaria hygrometrica (Ij.) Sibth. Formosa ; Prov. Shinchiku-cliö, Naroyama (Leg. Hisahiko Sasaoka ! 1. Xn. 1912.). Korea : in monte Rohô (Leg. Takenoshin Nakai ! 10. YII. 1914.). Species nova ad floram formosanam et koreanam. BRYACEAE. JSryujti argenteufii L. Formosa : Prov. Shinchiku-chö, Naroyama (Leg. Hisahiko Sasa.oka ! 1. xn. 1912.). Mhodohvyuni Wichurae (Broth.) Broth. Formosa : Prov. Taihoku-chô, Mentenzan (Leg. Yaichi Shimada ! 31. X. 1914.). Species nova ad floram formosanam. CONTRIBUÏIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 19 MNIACEAE. Milium Kiifoshii Sh. Oxamura. sp. nov. Habitat in saxis irroratis. Plantae robustiores, caespitosae, caespitibus late extentis, viridibus, in statu sicco molliusciilis nitidis. Caulis repens, usque ad 20 cm. longus, densissime fusco-radiculosus, irregulariter parce raraosus, den- «iuscule foliosus complanatus, sec- tione teres, c. 0.56-0.64 mm. diam., fascicule centrali c. 56 // diam. paucicellulari luteo vel luteo-fusco, reti intermedio hyaline vel lutescenti, cellulis hexagonis vel oblonge -hexagonis c. 40-56 [^ magnis, parietibus tenellis lutes- <3enti-fuscis, cellulis periphericis c. 3-4-seriatis sensim minoribus c. 10-18 fj- magnis paulum incrassatis fuscis ; ramis prostratis, c. 5 cm. longis, apice flageriformiter pro- ductis, simplicibus, densiuscule foliosis complanatis. Folia sicca leviter crispata, madida horizontalia et quasidisticha, haud decur- rentia, e basi valde angusta brevissima subito subcordato- saepe •ovato-oblonga, apice rotundato-obtusa, emarginata vel haud emar- ginata, nervo excurrente breviter apiculata, c. 5-9 mm. longa et c. 3.5-4.5 mm. lata, subplana, marginibus paulum undulatis, ubique limbatis, limbo lato lutescenti, inferne e seriebus 4-5, superiore 2-3 cellularum elongatum formate, fere e basi ad apicem dense et minute simplici-subserrulatis, serrulis obtusis ex unica série saepe A. X2 Fig. S. Mnitini Kiyoshii. A. Folia ( X 2) ; B. Apex foUi ( x 50) ; C. Sectio transversalis nervi ( x 50). 20 ART. 4. — SHUTAI OKAMURA : duabus Seriebus cellularum vesiciformium conipositis ; nervo valido, apice tenuiore, in sectione transversali biconvexo, dorso prominen- tiore, basi c. 0.32-0.36 mm. lato et c, 0.20-0.24 mm. crasso, cellulis ventralibus in medio 2-3-stratis minoribus incrassiusculis, duci- bus medianis numerosis laxioribus, fascicule comitum indistincte ducibus contiquo, fasciculo stereidarum valde minore pauci- cellulari (c. 6-8) luteo-fusco, cellulis dorsalibus c. 16 laxioribus ; cellulis laminalibus valde chlorophyllosis, laevibus, in seriebus obliquis dispositis, ad nervum oblonge -hexagonis, c. 70-100/^ longis et c. 35-50 [^ latis, ad limbum marginalem et apicem minoribus hexagonis, basilaribus infimis paulum elongatis subrectangulis. Caetera ignota. Hondo : Prov. Izu, iu inoute Araagi (Leg. KiYOSHi Fujn ! 18. YIII. 1914.). Species M. vesicato Besch. valde affinis, sed statura robustiore, foliis subcordato- vel ovato-oblongis, apice latioribus plerumque emarginatis faciliter dignoscenda. Nomen speciei in honoren Leg. Dom. Kiyoshi Fujii, Mnitiiii ligulifoliiini Card. Korea : iu monte Hötaizan [Ig. qè Uj] (Leg. Takenoshin Nakai ! 6. VILC. 1914.). Species nova ad floram koreanam. Distr. Hondo. llnium pu n etat Hin (L.) Hedw. Kyûshù : Prov. Ösmni, iu insula Yakusliima (Leg. Takanori Iwaki ! 23. IX. 1914.). Mniu^ii suhglohosnm Br. eur. Korea: in monte Gatokurei [sj-ft^] (Tjeg. Takenoshin Nakai! 5. CONTIIIBÜTIONES NOVAE AD ]7,0RAM BRYOrHYTON JAPONICAM. 21 Vn. 1914.) ; in monte Rohö (Leg. Takenoshin Nakai ! 11. Vn. 1914.). i Species nova ad floram koreanam. ' Mniuni fpichonianes Mitt. Formosa : Prov. Taihoku-cliö, Hakkosliö [/\ tp Œ] (Leg. Yaichi Shima- i da ! 20. VL 1914.). ! Species nova ad floram formosanam. RHIZOGONIACEAE. . Mhlisogoiiiuiii sphiifornie (L.) Bruch. Formosa : Prov. Tailiokn-cliö, Shabösan [i^^ ifs Uj] (Leg. Yaichi Shi- ^ MADA ! 31. Xn. 1913.). Species nova ad floram formosanam. AULACOMNIACEAE. Aulaconitiiiim turgidtini (Wahlenb.) Schwaeg. Korea ; in monte Rohö (Leg. Takenoshin Nakai ! 10. VIT. 1914.). Species nova ad floram koreanam. BARTRAMIACEAE. JPhilonotis palustris Mitt. Formosa : Urai (Leg. JuKiCHi Shiraga ! VII. 1911.). POLYTRICHACEAE. Pogonatuni arisanense Sh. Okamura. sp. nov. Habitat in terra. Plantae robustissimae, ferrugineo-virides sed sicca nigrescentes, hand nitidae, rigidiores. Caulis erectus, c. 12- 22 ART. 4. — SHUTAI OKAMÜRA : 15 cm. longus, cum foliis c. 1.5-2.5 cm. crassiis, strictns vel leviter flexuosns, rubiginosiis, basi dense fiiscescenti-radiculosus, simplex, dense foliosus, obtusus. Folia infima minuta, dein sensim majora, sicca adpressa crispata et circinato-involuta vel -convoluta et spiraliter hie illic torta, humida arcuate erecto-patentia vel patentia, e basi brevi (c. 2 mm. longo) semivaginante late ovali fascescenti et pellucido sensim lineari-lanceolata, apice breviter acuminata, c. 14-17 mm. longa, basi c. 2.0-2.5 mm. in memio c. 1.2 mm. lata, marginibus planis vel subplanis, e basi ad apicem dense (c. 80) serratis, serraturis basilares angustioribus lutescenti-fuscis ex unica Serie 1-3 cellularum formatis, in laminae parte infima minoribus autem ad apicem majoribus rubiginosis vulgo uni-cellula formatis ; nervo valido, in medio folii c. 0.16 mm. lato, cum apicem evanido, doro e medio ad apicem dense 2-4-seriate spinoso-serrato ; lamellae latitudinem folii totam occupantes, densissimae, sat numerosae (c. 70-80), humillimae c. 25-28 /^ altae, in sectione transversali ab uno Strato cellularum (2-3 rarius 4) constructae, rarius in basi vel medio duarum cellulae connatae, cellulis subaequalibus rarius ad margin majoribus, cellula marginali laevi haud incrassata quadrato- rotundata hemirotundata vel fere rotundata ; cellulis vaginalibus folii inferioribus elongate rectangulis c. 50-70,« longis et c. 15/^ latis, laminalibus minutis paulum incrassatis quadratis angusto quadratis vel subhexagonis c. 10-15/^ magnis, marginalibus quad- ratis c. 10/^ magnis. Inßorescentla dioica ; planta mascula ignota. Ramulus perlchaetialis caulis innovatione pseudolateralis. Bracteae 2)enchaetil externae erecto-patentiae, intimae suberectae, e basi longe (c. 3-4 mm.) semivaginante elliptica abrupte lineari-subulatae, apice attenuatae subcarinatae, c. 13-16 mm. longae, basi c. 1.6-2.0 mm. medio c. 0.56-0.70 mm. latae, marginibus fere e basi ad apicem dense serratis ; nervo valido, dorso superne 2 -seriate spinoso- CONTRIBUTIONES NOVAE AT) I'T.OEAM BRYOPHYTON JAPONICAM. 23 serrato. Vaghmla 6 mm. alta, longe cylindrico-conica, fusca, longe et hyaline pilosa. Seta suberecta, c. 3 cm. longa, rubra, fere stricta vel leviter flexuosula, sicca superne torta. Theca erecta, e collo in- distincte oblonga, c. 5 mm. longa et c. 2 mm. crassa, in sectione trans versali fere rotundata, leviter 8 -plicata, fusca, sicca deoperculata sub ore paulum contracta ; cellulis exothecii hexagonis quadrato-hexa- gonis Tel rectangulis c. 30-60 /^ longis et c. 20-30/^ latis, medio alte et recte papilliformi-prominen- tibus ; stomatibus nullis. Peristoma dentés 32, c. 0.4 mm. longi, e membrana basilari liumili (c. 0.08 mm.) lineari-oblongi, apice rotun- dato-obtusi, brnnnei margine liya- lini. Sporae minutissimae, c. 9-10/^ magnae, lutescentes vel oli- vaceae, laevissimae. Operculum et Calyptra mihi desunt. Formosa : in monte Arisan (Leg. BüNZö Hayata ! 6. IV. 1914.). Species cum P. contort o et P. atrovirentl plaecipue P. erythro- dontio comparanda, sed lamellis folii numerosis (c. 2.5-4.0 plo.) Fig. 9. Pogoiiatum arisanense. A. Planta fertilis (xl); B. Folium'(x5) ; C. Sectio transversalis lamellae folii ( x 500). 24 ART. 4. — SHÜTAI OKAMURA : facillime dignoscenda. Pogonatum grandifolium (Linde.) Jaeg. Korea : Prov. Heian-Hokudo, Hokn Jörnen [;lt ± M], Kokai-gun (Leg. Takenoshin Nakai ! 1. VIT. 1914.). Species nova ad floram koreanara. Pogonatutn inflexum Lindb. Formosa : Prov. Taihoku-cliö, Chiknsliiko [it 'ff -SB] (Leg. Yaichi Shi- MADA ! 3. I. 1914.) ; Prov. Taihoku-chö, Mentenzan [® 5^ |jj] (Leg. Yaichi Shimada ! 31. X. 1914.) ; iu monte Arisau (Leg. Bunzö Hayata ! 6. IV. 1914.). Pogonatiiin uruigeriini Schimp. Korea : Prov. Kankyo-Nandö, Daikori [;*: ^ M]- Cliôshin-guu (Leg. Takenoshin NaKAi ! 18. Yn. 1914.). Species nova ad floram koreanara. Polytrichuin alpinum L. Korea : in monte Roliô (Leg. Takenoshin Nakai ! 11. YII. 1914.). Species nova ad floram koreanam. Polytrichuin commune L. Sachalin : Sicca (Leg. Jüsuke Susuki ! 1. Y. 1913.). Korea : Prov. Kankyo-Nandö, Hakutokurei [â W. ^], Kösan-gun [ep Oj fp] (Leg. Takeno- shin Nakai ! 10. YII. 1914.) ; in monte Rohö (Leg. Takenoshin Nakai ! 4. Yin. 1914). Species nova ad floram koreanam. CONTEIBUTIONES NOVAE AD FLoRAM BRYOPHYTON JAPONICAM. 25 HEDWIGIACEAE. Medtvigia albicans (Web.) Lindb. Korea : Prov. Heiau-HoknclO, in monte Hiraihô [^ ?}i iii$] (Leg. Take- NOSHINiNAK-\I ! 9. VI. 1914.). CLIMACIACEAE. Climaciuni japonicum Lindb. Korea : Prov. Heian-Hokudô, in monte Hiraihô (Leg. Takenoshin Nakai! 9.-4YL 1914.). I*leurorHopsis viithenica (Weinm.) Kindb. Korea : in monte Gatoknrei (Leg. Takenoshin Nakai ! 5. VIE. 1914.) ; in monte Roliö (Leg. Takenoshin Nakai ! 11. VII. 1914.). Species nova ad floram koreanam. CRYPHAEACEAE. PilotricJiojysis dentata (Mitt.) Besch. Formosa : Prov. Shinchikn-chô ; ßitösan [^ ff lü] (Leg. HiSAHiKO Sasa- OKA ! 2. xn. 1912.J. LEUCODONTACEAE. Leucodon perdeiyeudeïis Sh. Okamura. sp. nov. Habitat in rarais arborum (Picea ajanensis et Abies nephrolepis etc.) sylvaticarum in alpinis. Plantae graciles, caespitosae, caespiti- bus viridibns vel fnscescenti-viridibus, nitidiusculis, rigidiusculis, 26 ART. 4. — SHÜTAI OKAMURA 7-^1 densiusculis. Caulis primarius repens, filiformis ; secundarius longe dépendons, snbstrictus, usque ad 10-20 cm. longus, irregulariter et laxe subpinnatim ramosus, sectione ellipticus, c. 0.24-0.28x0.16- 0.2 mm. magnus, fasciculo central! nullo, reti centrali hyaline, ce- Uulis hexagonis c. 15-20// magnis, tenellis, cellulis periphericis c. ö-4-seriatis minoribus rubiginosis valde incrassatis ; ramis pendulis, c. 1-6 cm. longis, apice longe vel brève attenuatis, simplicibus vel parce (1-5) ramulosis ; ramulis c. 1 cm. longis, patentibus, apice attenuatis vel obtusis ; ramis ramulisque sicca apice substrictis arcuatis vel hamulosis sed haud circinato-incurvis ; caulibus secundariis ramis ramulisque densiuscule foliosis teretibus ; caulis secundarius floriferous saepe brevis, simplex vel parce ramosus. Folia sicca adpressa argute longitudinaliter plicata, madida erecto-patentia, valde concava, profunde plicata, basi breviter decur- rentia, oblongo-lanceolata apice breviter attonuata vel acuta, summo apice cana- liculata, c. 1.2-1.9 mm. longa et c. 0.5-0.8 mm. lata, marginibus late planis, apice minute serrulatis ; nervo nullo ; cellulis laevissimis, valde chlorophyllosis, linearibus, c. 30-56// longis et c. 3-4// latis, marginalibus basilaribus pluriseriatis (c. 12) quadratis incrassatis c. 9-12/^- magnis, basilaribus infimis per totam latitudinem folii lutescenti-fuscis. Inßorescentia dioica, utriusque sexus in cauligenis vel ramuligenis ; flores masculi geminiformes ; folia perigonialia concava, ovali-oblonga, apice cuspidata, enervia ; antheridia c. 10, paraphysibus numerosis. Bracteae perichaetü X20 Fig. 10. Leucodon perdepéndens. A. Folia ( X 20) ; B. Sporangia (xlO). CONTRIBUTIONES NOVAE AD IX0RA3I BRYOPHYTON JAPONICAM. 27 longitudine fere setae, intimae e basi alte vaginante breviter acuminatae, erectae apice, erecto-patentiae, integerrimae, énerves, c. 3.6 mm. longae, baud plicatae ; celluHs linearibus, valde in- crassatis, marginalibns basilaribus rectangulis. Vaglnula cylindrica, c. 1.2 mm. alta, fuliginosa. Seta erecta, stricta, c. 3 mm. longa, basi rubiginosa dein lutea, laevis, sicca leviter torta. Theca erecta, symmetrica, oblonga vel elliptico-oblonga, c. 1.2 mm. longa et c. 0.56 mm. crassa, sicca deoperculata snb ore panlnm contracta, lutea vel luteo-fusca, laevis, brevicollis ; cellulis exothecii irregulariter qnadratis hexagonis vel rectangulis, baud incrassatis, ad orificium in seriebus 4-5 minutis anguste qnadratis ; stomatibus null is. Am'mlus duplex, c. 40 !>■ altus, basi Intens apice hyalinus, deciduus. Exostomli dentés lineari-lanceolati, basi connanti, c. 0.28 mm. longi et basi c. 40 p- lati, linea media flexuosula sed indistincta baud perforata, strato dorsali lutescenti basi laevi dein dense et grosse papilloso, ventrali dense (c. 15) et humile lamelloso ; endostomium flavidulum, grosse papillosum ; corona basilaris c. 70 p- alta ; pro- cessus nulli. Sporne c. 15^ magnae, ochraceae, laeves. Oper- culum ignotum. Calyptra cucullata, c. 2.2 mm. longa, lutescenti- viridescens apice brunnoa, laevis. Korea : Prov. Heian-Hokudô, in monte Aibutsusan [^ f?ï Uj] (1320 m.) (Leg. Takenoshin Nakai ! 1. VIT. 1914.). Species L. pendulo Lindb. comparanda, sed seta laevis, exo- stomii dentés basi leavi dien grosse papillosi longe diversa. Nomen speciei ab caule secundario perdependenti. NECKERACEAE. I*terohvijoj)sis ciicullatifolia Sh. Okamura. sp. nov. Habitat in truncis arborum. Plantae robustiusculae, lutescenti- 28 ART. 4. — SHUTAI OKAMURA : viridescentes, nitidiusculae, moUiusculae. Caulis primarius repens, hie illic fasciculatim fusco-radiculosus ; secundarius erectus, c. 3-4 cm. longus, simplex vel parce ramosus, sectione teres, c. 0.3-0.35 mm. crassus, fasciculo centrali nullo, reti centrali luteo, cellulis hexagonis vel oblongo-liexagonis, incrassiusculis c. 15-20/^ magnis, cellulis inde ad snperficiem paulum minoribus incrassatioribus rubiginosis ; ramis erecto-patentis, c. 1-2 cm. longis, obtusis vel subflagelliformiter attenuatis ; caulibus secundariis ramisque dense et turgide foliosis teretibus. Folia sicca subimbricata, madida erecto-patentia, baud decurrentia, cymbiformi-con- cava, ovato-oblonga, raptim in acu- men angustum strictum vel leviter reciu'vum plus minus longum (c. 0.4-0.5 mm.) subulatum constricta, c. 2.4-2.6 mm. longa et c. 1.1-1.3-1.5 min. lata, marginibus superne valde inliexis, in subula minute serrulatis, caeterum integerrimis ; nervo sim- plici, tenui, c. 0.7-1.6 mm. longo, saepe furcato ; cellulis chlorophyl- losis, linearibus, parietibus incrassatis et plus minus porosis, in medio folii paulum porosis c. 70-100/^ longis et c. 7/^ latis, supe- rioribus argute porosis c. 40-60 /^ longis et c. 7 /^ latis, basilari- bus infimis fuscis valde incrassatis brevioribus latioribus c. 15/^ latis inter sese valde porosis, alaribus distinctis fuscis oblongis inter sese valde porosis c. 20-40 j« longis et c. 15-18/^ latis. Caetera ignota. Formosa : Prov. Sinchiku-cliö, Naroyama (Leg. Hjsahiko Sasaoka ! 1. Xn. 1912.). Species cun P. crassicauli (C. Müll.) Fleisch, et P. crassiusculo A. yi5 /**;/. 11. Pterobryopsis ciicviUatifolia. A. Folia. (xl5). CONTBIBUTIONES NOVAE AD l-LORAM P.RYOPHYTON JAPONICAM. 29 (Card.) Broth, comparanda, ab hic statura minoriore, foliis in acumen 1/3 longis, ah illia folioriim forma dignoscenda. Aërohryopsis JParisii (Card.) Broth. Engl, und Prant. uat. Pflauzenf, j)- 820. (1906). [nomeii solum comb. nov.]. Syn. Aërohryum Ferriei Broth, in Bull, de I'Herb. Boissier. 2'™ Ser. n. p. 926. (1902). [uomen solum !]. Syn. Meteorkim Parisii Card. Reih. Bot. Centralblatt. Band. XIX. Abt. n. Heft. 1. p. 121. (1905). [Ster. !]. Habitat in truncis arbornm. Pkintae robustiusculae, caespito- sae, caespitibus Intescentibus vol lutescenti-viridibus, nitidiusculis, mollibus. Caulis primarius elongatus repens, hie illic fasciculatim fuscescenti-radiculosus ; secundarius irregulariter pinnatus, longe dépendons, substrictns vel leviter flexuosus, usque ad 10-20 cm. longus, laxe inordinate pinnatim ramosus, sectione teres vel saepe oblongus, c. 0.28-0.36x0,24-0.28 mm. magnus, fasciculo centrali nullo, reti centrali hyaline vel lutescenti, cellulis hexagonis c. 15-20 />« magnis, parietibus incrassiusculis, cellulis periphericis c. 5-6-seriatis minoribus valde incrassatis luteis vol fuscis ; ramis patulis vel patentibus, c. 1-3 cm. longis, strictis vel leviter flexuo- sis, plerumque simplicibus rarius parce ramulosis, obtusis vel acutis ; caulibus ramis ramulisque dense foliosis subcomplanatis. Folia sicca laxe adpressa rugulosa mollia, madida patentia vel erecto- patentia, valde concava, ovata vel oblonga, in acumen elongatum undulatum flexuosum sensim attenuata apice capillare sat subito constricta, c. 2.0-3.2-3.5 mm. longa et c. 0.7-1.2 mm. lata, mar- ginibus planis vel subplanis, superne undulatis, ubique minutissime serrulatis ; nervo tenui, ad basin acuminis evanido, in sectione transversali plane biconvexo, basi c. 42 /j- lato et c. 15 (j- crasso, e 3-stratis ceUularum incrassatarum uniformium composite ; cellulis 30 ART. 4. — SHUTAI OKAMÜKA : FifJ. 12, Aërobryopsis Parisii. A. Folia ( X 15) ; R. Theca ( x 5). chlorophyllosis, parietibus omnibus incrassatis, in medio folii rhom- beis vel sublineari-rhomboideis, dorso papilla singula medio notatis, c. 20-40 !i longis et c. 7 /^ latis, basilaribus linearibus c. 40- 56 /^ longis (^t c. 7 !'■ latis, in- fimis latioribus c. 7-1) n sinua- tis inter sese porosis laevibus, alaribus numerosis brevioribus oblongis rectangulis ovatisque c. 20-30/^ longis et c. ISa' latis lutescentibus laevibus, superioribus sublineari-rhom- boideis c. 20-40/^ longis et c. 5-7 /^ latis laevibus. Inßores- centia dioica ; flores masculi ignota ; flores feminei in caule vel ramis. Ramulus pevlchaetialis baud radiculosus. Bracteae perichaetli intimae basi haud vaginatae oblongae concavae, in acumen recurvum vel erectum elongatum filiforme minute serrulatum subito attenuatae, c. 2.4-2.8 mm. longae et basi c. 0.45 mm. latae ; nervo tenui, infra apicem evanido ; cellulis linearibus, inter sese porosis, basilaribus laxioribus, omnibus laevissimis. Vaginula cylindrica, c. 1.6-1.8 mm. longa, lutescenti- viridescens apice fusca. Seta saepe gemella, c. 10-15-20 mm. longa, stricta vel leviter flexuosula, rubra, inferne laevis superne scabra, sicca haud torta. Theca suberecta vel arcuata, asymmetrica, leviter gibbosa, e collo distincte conico cylindrica, ad orificium paulum angusta, sicca deoperculata sub ore paulum contracta, haud plicata, c. 2.5-3.0 mm. longa (collo c. 0.4 mm.) et c. 0.8 mm. crassa, fusca, laevis ; cellulis exothecii quadratis vel rectangulis c. 25-56;" longis et c. 20-25/^ latis, parietibus incrassiusculis, ad CONTßlBÜTIONES NOVAE AD FT.ORAM BRYOPHYTON JAPONICAM. 31 orificium in seriebus c. 5 minoribus quadratis c. 15-20/^ magnis; stomatibus in coUo numerosis phaneroporis. Annulus duplex, c. 30-35 n altus, ex unica vel duplici série cellnlarnm, fusons, deciduus. Exostomil dentés basi connati, lineari-subulati, c. 0.56 mm. longi et basi c. 56 « lati, albescentes, dorso minute papillosi, ventro altius- cule c. 25-lamellosi ; endostomium albicans vel lutescens ; corona basilaris c. 100-110/^ alta, minutissime papulosa vel fere laevis ; processus dentium longitudinis, lineari, dense et minute papillosi, hand carinati, rimoso-perforati ; cilia nulla. Sporae c. 14-16« magnae, luteae, paulum scabrae. Operculum e basi conica sensim attenuate rostratum, c. 1.0-1.2 mm. longum et basi 0.48-0.56 mm. diam., rostro paulum curvato. Calyptra ignota. Formosa : Knshak et Tamusui (Leg. Faueie ! 1903. ster.) ; Prov. TaihokH-cho, Ural (Leg. Jükichi SmBAGA ! VlI. 1911. ster.) ; Prov. Sliin- chiku-cliö, Naroyama (Leg. HiSAHiKO Sasaoka ! 1. Xu, 1912. fr.). Distr. KyiTshü : Prov. Ösumi ; Amami-Öslnrna (Leg. Faurie ! VI. 1900. ster.). A'êrobryopsis assimilis (Card.) Par. Formosa : Prov. 8hincliiku-cliô, Naroyama (Leg. Hisahiko Sasaoka ! 1. Xn. 1912.). Aërobryopsis sttbdivergens (Broth.) Broth. Formosa : Prow Taihoku-chô, Mentenzan (Log. Yaichi Shimada ! 31. X. 1914.). 3Ieteot'iopsis ancistvoides (Ren. et Card.) Broth. Engl. und Prant. nat. Pfranzenf. p. 826. (1906). 8yn. Meteorium aucistroides Ren. et Card, in Bull. Soc. voj. Belg. I. p. 72. (1895) ; Meteorium liymcdayense Par. lud. Br. Editio secuuda Vol. 111. p. 229. (1905). 32 ART. 4. — SHUTAI OKAMURA : Habitat in truncis arboriim. Plmitae robustiusculae, caespitosae, caespitibus lutescenti-viridescentibus, nitidiuscnlis, moUiusculis. Caulis Primarius elongatns, repens, divisus, hie illic fasciculatim fuscescenti-radiculosus ; secundariiis longe dépendons, leviter flexu- osus, c. 8-20 cm. longus, dense pinnatim ramosus, sectione teres, c. 0.32 mm. diam., fasciculo centrali nullo, reti centrali hyalino, cellulis hexagonis tenellis c. 14/^ magnis, cellulis periphericis c. 5-8- seriatis sensim minoribus valde incrassatis Inteis vel fuscescentibus ; ramis patentibns vel horizontalis, c. 6-30 mm. longis, strictis leviter curvatis vel panlum flexuosis, hand hainnlosis, simplicibns saepe parce (1-4) ramulosis, obtnsis ; caiilibus ramis ramulisque dense squarrose foliosis teretibns. FoUa sicca imbricata vel leviter cris- pata, e basi breviter decurrente suberecta semiamplexicauli late ovata vel i'otundato-ovata, apice sensim brève vel longe subulato- attenuata, subula subcanaliculata recurvo-squarrosa, c. 1.6-2.0 mm. longa et c. 0.8-1.0 mm. lata, concava, hand plicata, marginibus ubique praecipue ad apicem argute serratis vel serrulatis, apice saepe undulatis ; nervo tenui, infra apicem folii evanido, in sectione transversali plano-convexo, dorso paulum prominenti, basi c. 40;« lato et c. 20/^ crasso, e 3-stratis cellularum incrassatarum uni- formium composite ; cellulis chlorophyUosis, incrassatis, utrobique facie papilla singula medio notatis sed hand opacis, longe rhom- boideis vel breviter linearibus, c. 28-40/^ longis et c. 3-4 /^ latis, basilaribus longioribus latioribus c. 40-60 f^ longis et c. 7-10 r- latis laevibus inter sese porosis, in alis saepe paulum concavis distinctis (rarius indistinctis) numerosis (c. 20-40) quadratis vel oblongis c. 14-28/^ longis et c. 10-15;^ latis baud porosis vel parce porosis hyalinis vel fuscescentibus. Inßorescentia dioica ; flores masculi ignoti ; flores feminei in ramis vel caule. Ramulus peri- chaetialis hand radiculosus. Bracteae perichaetii intimae deltoideo- CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 33 A. ^^ x20 Fig. 13. Meteoriopsis ancistroides. A. Folia ( X 20) ; B. Sporangia ( x 5). oblongae, in acumen subito erectuni saepe flexuosura elon- gatum angustum subpiliforme sensim attenuatae, c. 1.6- 1.8 mm. longae et basi c. 0.56-0.7 mm. latae, margini- bus in acumen argute serratis ; nervo saepe indistincte, me- dium evanido ; cellulis laevi- bus, linearibus, c. 40-80 /^ longis et c. 7-10/^ latis. Vag inula cylindrica, c. 1.2 mm. longa et c. 0.48 mm. crassa, lutescens apice fusca, valde pilosa. Seta c. 1.2-1.8 mm. alta et c. 0.16-0.24 mm. crassa, apice crassiore, erecta, fusca, laevis. TJieca e perichaetio exserta, erecta, symmetrica, ovato-oblonga, sicca deoperculata sub ore paulum contracta, baud plicata, cum coUo c. 2.5-2.7 mm. longa (collo c. 0.5mm.) et c. 1.1— 1.2mm. crassa, fusca, laevis; cellulis exotheci hexagonis vel oblongo-hexagonis, c. 35-50 /^ longis et c. 20-30 /^ latis, ad orificium in seriebus c. 5 minoribus hexagonis c. 15 ;" magnis ; stomatibus nuUis. Exostomii dentés basi paulum remoti, lineari-lanceolati, c. 0.64 mm. longi et basi c. 0.084 mm. lati, strato dorsali lutescenti, e basi ad medium transversim striatulo inde laevi, linea media flexuosula, strato ventrali latiore, densissime (c. 35-40) et alte lamelloso ; endostomium hyalinum vel fuscescens, minute papillosum ; corona basilaris c. 0.13 mm. alta, horride plicata ; processas c. 0.48 mm. alti, carinati in carina auguste perforati, minute papillosi ; cilia nulla. Sporae c. 30-35 ,« magnae, ferrugineae, laeves. Operculum e basi conica rostratum, c. 0.9 34 AET. 4. SHÜTAI OKAMURA mm. longum et c. 0.7 mm. diam., rostro longiusculo recto. Calyptra ignota. Formosa : Prov. Shincliiku-chö, Naroynnia (Leg. Hisahiko Sasaoka ! 1. XII. 1912.). Distr. Sikkim-Himalaya ; Bhotaii. lleteoriella cuspidata 8h. Okamura. sp. nov. Habitat in truncis arborum. Plantae robustae, caespitosae, caespitibus lutescenti-viridescentibus, nitidinsculis, rigidiusculis. Caulis Primarius elongatus, repens, divisus, hie illie dense fascicula- tim fusco-radiculosus ; secundarius longe dépendons, leviter flexuo- sus, c. 6-28 cm. longus, apice obtusus rarius flagelb' for miter at- tenuatus, dense vol saepe laxe regulariter pinnatim ramosus, sec- tione ellipticus, c. 0.36-0.48-0.72x0.24-0.32-0.45 mm. magnus, fasciculo centrali nuUo, reti centrali hyalino vel lutescenti-rubiginoso, cellulis liexagonis vel oblongo-hexagonis c. 15-20-30/^ magnis, parietibus tenellis vel incrassiusculis, cellulis peripliericis c. 5-7- seriatis sensim minoribus valde incrassatis rubiginosis ; ramis 1-3 cm. saepe 7 cm. longis, patenti- bus horizontalis vel pendulis, strictis curvatulis vel leviter flexuosis baud hamulosis, obtusis rarius flagelli- formiter attenuatis, simplicibus vel dense vel laxe pinnatim ramulosis ; ramulis c. 5-10 mm. longis, obtusis rarius flagelliformiter attenuatis ; caulibus secundariis ramis ramulis- que dense et turgide foliosis, tereti- bus vel leviter complanatis, cum foliis c. 2.5-3.0 mm. crassis. Folia sicca vix mutata, e basi cordata erecta aurieulata (auriculis parce Fig. 14. Meteoriella cuspidata. A. Folia (xl5.) CONTKIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 35 circinatis caulem amploxis) late ovato-oblonga, apice breviter cuspidata (cuspide erecto vel leviter recur vo), valde concava, c. 2.4-2.6 mm. longa et c. 1.5-1.6 mm. lata, marginibus ubique integerrimis vel panlum crenulatis, ad liasin acnmiiiis inflexis ; nervis binis, aequalibus vel inaequalibus (saepe nervus unus in- distinctus), c. 0.32-0.8 mm. longis, in sectione transversal! basi c. 40-56;" latis et c. 20 y- crassis, e 2-stratis cellularum incrassatarum compositis ; cellulis linearibiis, laevibus, parietibns ubique incrassatis minute porosis, in medio et apice c. 30-50 !>■ longis et c. 7 /^ latis, basilaribus c. 56-75/^ longis et c. 7/-^ latis, infimis brevioribus et latioribus rectangulis c. 30-35/^ longis et c. 10-14/^ latis fuscis vel rubiginosis, alaribus nullis. Caetera ignota. Formosa : Prov. Sliincliiku-clio, Naroyama (Leg. Hisahiko Sasaoka ! 1. Xn. 1912.); Prov. Taihoku-chô, Urai (Leg. Jukichi Shiraga ! Vu. 1911.). Species distinctissima, a congeneribus foliis apice breviter cus- pidatis, margine integerrimis vel subintegerrimis faciliter dignoscenda. Nomen speciei ab apice cuspidato folii. Jleteoriella sohcta (Mitt.) Sh. Okamura. Jour. Coll. Sei. Imijer. Univ. Tokyo. Vol. XXXVI. Art. 7. p. 18. (1915.) Syn. Meteorium solutwii Mitt. Musci Ind. orient. [Jour, the proce. Lin. Soc. Supp. Bot. Vol. I. p. 88. (1859)]. Habitat in rupibus Plantae robustae, fusco-nigrescentes apice lutescenti-virides, rigidiusculae, nitidiusculae, caespitosae. Caulis Primarius repens, elongatus, usque ad 13 cm. longus, hie illic fas- ciculatim fusco-radiculosus, divisus ; secundarius pendulus, usque ad 5 cm. longus, strictus vel leviter flexuosus, densiuscule et re- gulariter pinnatim ramosus, sectione elHpticus, c. 0.28-0.32x0.20- 0.24 mm. magnus, fasciculo centrali nullo, reti centrali subrubigi- 36 ART. 4. SHUTAI OKAMUEA : Fiff. 15. Meteoriella soluta. A. FoUa (x20). noso, cellulis oblongo-liexagonis paulum incrassatis c. 20-28 x 15- 20 /^ magnis, cellulis periphericis c. 3-4-seriatis minoribus hexagonis vel oblongo-hexagonis c. l~10fi magnis valde incrassatis rubiginosis ; ramis c. 1-2.5 cm. longis, patnlis, densiuscule foliosis teretibus, apice obtusis vel breviter attenuatis, simplicibus vel 1-o-ramulosis ; ramnlis usque ad 5 mm. longis. Folia sicca vix mutata, e basi cordata erecta auriculata (auriculis parce circinatis caulem amplexis) ovata elliptica vel subrotundata, apice subu- lato-acuminata divaricata vel recurvo-squarrosa, c. 2.0-2.4 mm. longa et c. 1.0-1.1 mm. lata, valde concava, marginibus sub- integerrimis vel minutissime serrulatis ; nervis binis, inaequalibus, usque ad medimn folii evanidis, in sectione transversali basi e 2-stratis cellularum uniformium incrassatarum compositis ; cellulis laevissimis, linearibus, parietibus incrassatis minute porosis, c. 42-56 II longis et c. 7 /^ latis, superioribus c. 28-40 /^- longis et c. 4-5, a latis, basilaribus abbreviatis latioribus c. 15-20/^ longis et c. 15/^ latis fuscis. Houdo : Prov. Ise, in monte Gozaislio, Mie-gun (Leg. Kichitaro Mu- EATA ! 17. X. 1914.). Species nova ad floram japonicam. Distr. Sikkim. Meteoriella Kiitôi (Sh. Okamuba) 8h. Okamuka. Com. nov. Syn. Meteoriella solufa (Mitt.) Sh. Okamuea. var. Kufôi Sh. Okamuea. Jom-. CoU. Sei. Imper. Univ. Tokyo. Yol. XXXYI. Art. 7. p. 18, (1915.). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 37 Conspectus Specierum Meteoriellarum. Folia apice ciispiclata leviter recnrva vel erecta, plautae robustae. M. cuspidafa. Folia apice subulato-attenuata recurvo-squarrosa 1. Plantae robustae, folia apice integerrima vel minutissime serrulata 1. • M. soluta. Plautae teuellae, folia apice majore deutata vel retro-serrata . . 31. Kutbi. Trachypus Jmniilis Lindb. Formosa : Prov, Tailioku-chô, Urai (Leg. Jukichi Shiraga ! Vu. 1911.). Species nova ad floram formosanam. Huuantocladiiim lorifovnie (Br. Jav.) Fleisch. Habitat in trnncis arborum. Plantae robustae, caespitosae, caespitibus Intescenti-viridibus, nitidiusculis, moUiusculis, Caulis Primarius repens, elongatus, hie illic dense et fasciculatim fusco- radiculosus, saepe stoloniferus ; secundarius remotus (c. 7-10 mm. remotus), dépendons, strictus, c. 6-12 cm. longus, apice longe attenuatus saepe flagellaris, basi stipitatus inde dense regulariter pinnatim ramosus, stipite c. 1-3 cm. longo, e stipite flagella par- vifolia et fasciculatim radiculosa emittentes, sectione teres vel oblongus, c. 0.45-0.64x0.35-0.56 mm. magnus, fascicule centraU nuUo, reti centrali hyaline, cellulis hexagonis vel oblongo-hexagonis c. 20-30/^ magnis teneUis, cellulis periphericis c. 5-9-seriatis sensim minoribus crassioribus lutescentibus vel fuscis ; ramis c. 2-3 cm. longis, patentibus, apice obtusis vel longe attenuatis saepe flagellaribus, plerumque regulariter pinnatim ramulosis ; ramulis c. 5-10 mm. longis, apice obtusis vel longe attenuatis saepe flagelliformibus ; caulibus secundariis ramis ramulisque densiuscule foliosis, valde 38 ART. 4. — SHUTAI OKAMURA complanatis, cum foliis c. 2-3-4 mm. latis. ParajyhyUia nulla. Folia unimorplia, octosticlia, sicca vix mutata transverse rugulosa, dorsalia subimbricata, lateralia erecto-patentia, paulum asymmetrica, breviter decurrentia, e basi cordato-ovata longe lingulata, apice latiora et acuta vel cuspidato-acuta, concaviuscula, baud longitudi- naliter plicata, c. 2.4-2.8-3.0 mm. longa et basi c. 1.0-1.2 mm. in medio c. 0.04 mm. lata, marginibus basi uno latere late inflexo, inferne minute serrulatis, apice grosse serratis ; nervo valide, infra summum apicem folii evanido, in sectione transversali piano -convexo, dorso paulum prominenti, basi c. 40-50 ft lato et c. 25 /^ crasso, e 3-stratis cellularum incrassatarum uniformium composite ; cellulis chlorophyllosis, laevissimis, pellucidis, parietibus omnino incrassius- culis c. 1.0-1.5 /-« latis, luminibus e medio ad apicem folii 4-6- angulato-rotundatis vel ovalis c. 7-10-15/^ magnis, marginalibus e medio ad infra apicein usque ad 5-seriatis elongatis rhomboideis vel breviter linearibus c. 20-40 /^ longis sed unica série externa plerumque breviore, basilaribus basin versus sen- sim longioribus Lite rhomboideis inde linearibus c. 30-70/^ longis et c. 7-9/^ latis inter sese porosis, angularibus indistinctis sed saepe quadratis vel breviter rectangulis minoribus ; folia ramea et ramulia minora c. 1.6-1.8 mm. longa, basi c. 0.8-0.9 mm. in medio folii c. 0.50-0.64 mm. lata. Caetera iç-nota. Fig. 16. Hymantocladium loriforme. A. Folia (xl5). Formosa: Prov. Taihoku-clio, Urai (Leg. Jukichi ShiraCxA ! YII. 1911.). Species nova ad floram japonicam. CONTRIBUTIONES NO"\'AE AD FLOEAM BEYOPHYTON JAPONICAM. 39 Distr. Celebes, ins. Ceram, Banca, Java, ins. Fidji, Samoa, Borneo, New Guinea, Philippines etc. Neckevojisis iJseudonitidiila Sh. Okamuea. sp. nov. Habitat in truncis arborum. Plantae valde graciles, caespitosae, caespitibus lutescentibus vol lutescenti-viridibus, nitidiuscnlis, moUi- usculis. Caulis primarius repens, elongatus, filiformis, c. 4-6 cm. longus, hie illic dense et fascicnlatim fnsco-radiculosus ; secundarius densiusculiis, dependens, strictus vel fere strictus, c. 2-7 cm. longus, apice flagelliformis, simplex vel laxiuscule vel densiuscule et ir- regulariter pinnatim ramosus, sectione teres vel oblongus, c. 0.12- 0.24x0.12 mm. magnus, fasciculo centrali nuUo, reti centrali hya- line, cellulis oblongo-hexagonis vel hexagonis c. lO-loA« magnis, parietibus tenellis vel incrassinsculis, cellulis periphericis c. 4-6- seriatis c. 10-15/^ magnis valde incrassatis lutescentibus vel aureis ; ramis c. 7-25 mm. longis, patentibus vel dependentibus, fere strictis, apice obtusis vel longe attenuatis et saepe flagelliformiter productis, simplicibus rarius parce (1-2) brevi- ter ramulosis ; caulibus secundariis ramisque densiuscule foliosis, valde complanatis, cum foliis c. 1.0-1.3 mm. latis, saepe flagelliferis, flagellis elongatis microphyllinis. Paraphyllia nulla. Folia sicca immutata, hand rugulosa, quadristicha sed ut videntur distiche complanata, erecto-patentia vel divaricato-patentia, asymmetrica, breviter decurrentia, obovata vel rotundato-obovata, apice rotundato-obtusa, c. 0.64-0.8-0.9 mm. longa et c. 0.48-0.58-0.64 mm. lata, subplana, marginibus e basi ad X300 Fi(J' 17. Neckeropsis pseudonitidula. A. folia ( X 20) ; B. Celliilae e parte apicis folii ( x 300). 40 ART. 4. — SHUTAI OKAMURA : medium uno latere auguste inflexo, basi crenulatis, apice crenulatis vel minute serrulatis, baud limbatis ; nervo validiusculo, ad 5/10-7/10 evanido (c. 0.4-0.64 mm. longo.), in sectione transversali fere piano, basi c. 30/^ lato et c. 15/^ crasso, e 2-stratis cellularum subincrassatarum uniformium composite ; cellulis chlorophyllosis, laevissimis, parietibus omnino incrassiusculis c. 2-3 /^ latis, luminibus superioribus rotundatis quadrato-rotundatis vel breviter oblongis, c. 7-9-15/^ longis, e medio folii basin versus sensim longioribus, late ovalis rhombeis vel breviter rhomboideis, c. 15-20 /« longis et c. 7-9// latis, basilaribus ad nervum linearibus c. 18-30/^ longis et c. 5-7 A« latis hand porosis, marginalibus quadratis breviter rectangulis vel ovalis, angularibus indistinctis. Caetera ignota. Formosa : Prov. Tailiokii-cliô, Ural (Leg. Jükichi Shiraga ! Vu. 1911.). Species N. nitidula Mitt, valde proxima, sed statura graciliore, foliis apice rotundato-obtusis baud acutis, marginibus minute serrulatis, nervo longiore et crassiore, cellulis superioribus luminali- bus pier unique rotundatis vel rotundato- quadratis jam abunde diversa. Nomen speciei ab fecie ut N. nitidula Mitt. Neckeva idunioana Sh. Okamura. sp. nov. Habitat in truncis arborum. Plantae robustiusculae, saespi- tosae, caespitibus laete viridibus inferne stramineois, nitidiusculis, molliusculis. Caulis priraarius elongatus, repens, hie illic fasciculatim fusco-radiculosus ; secundarius dépendons, strictus, c. 4-8 cm. longus, dense pinnatim ramosus, apice obtusus, sectione oblongus, c. 0.20-0.30x0.16x0.2 mm. magnus, fasciculo centrali nullo, reti centrali hyaline, cellulis hexagonis c. 10-15 // magnis, parietibus paulum incrassatis, cellulis periphericis c. 3-6-seriatis minoribus incrassatis luteo-fuscis ; ramis c. 5-25 mm. longis, CONTßlBÜTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 41 simplicibus vel pinnatim ramulosis, patentibus, strictis ; ramulis c. 5-7 mm. longis, apice obtusis saepe flagellif or miter productis ; cau- libus secundariis ramis ramulisque dense foliosis, plus minus com- planatis, cum foliis c. 2.5-3.0 mm. latis. Paraphyllia nulla. Folia sicca immutata, haud transverse rugulosa, dorsalia imbricata, sym- metrica, apice obtuse acuta, lateralia erecto-patentia vel patentia, paulum asymmetrica, breviter decurrentia, ovato- vel late ovato- oblonga, apice breviter et auguste vel latiuscule acuta, c. 1.6-1.8 mm. longa et c. 0.7-0.96 mm. lata, concava, haud plicata, mar- ginibus e basi ad medium uno latere auguste inflexo, e medio ad apicem serrulatis vel fere integerrimis ; nervo tenui, medium vel ultra medium folii evanido, in sectione transversali piano -convexo, dorso paulum prominenti, basi c. öß fi lato et c. 25/^ crasso, e 3-stratis cellularum incrassatarum uniformium composite ; cellulis valde chloropliyllosis, laevibus, parietibus incrassatis, superioribus rhombeis vel oblongis c. 10-15-20^ longis et c. 7-8/^ latis, basi- laribus linea'ri-rhomboi- deis vel linearibus c. 25-50 <" longis et c. 7 p- latis parce porosis, mar- ginalibus brevioribus, in alis concaviusculis ob - lonoris vel rectano;ulis c. 15-28 /^ longis et c. B. xio a; C.XIO X20 I i(/. 1 8. Keckera idumoana. 9-11 /J- latis parce porosis a. Folla (x20); B. Bractea perichaetialia intima ( x 10) ; T , „ . -r ,T C- Sporangia ( x 10) ; T>. Calyiîtra ( x 20). luteo-fuscis. Inßorescea- t'ia dioica (Acribus masculis haud visis dioica videtur) ; flores feminei numerosi in caulibus secundariis vel ramis. Ramulus peri- cJiaetialis haud radiculosus. Bracteae perich aetil intimae basi alte vaginatae et oblongo-lanceolatae, apice auguste attenuatae, c. 42 ART. 4. SHUTAI OKAMURA : 3.2-3.7 mm. longae et c. 0.8 mm. latae, e medio ad apicem minute serrulatae ; nervo medium evanido ; cellulis linearibus c. 30-70 [j- longis et c. 7/'- latis, basilaribus rectangulis c. 25-50," longis et c. 12 fj- latis, angularibus 4-6-gonis. Vaglmda cylindrica, c. 0.8 mm. longa, fusca ; paraphysibus numerosis hyalinis. Seta c. 0.4-0.5 mm. longa et c. 0.18-0.2 mm. crassa, erecta, fusca, laevis. T/ieca immersa, oblonga, symmetrica, erecta, c. 1.2-1.6 mm. longa et c. 0.95 mm. crassa, fusca, laevis ; cellulis exothecii liexagonis vel rectanguli-hexagonis plus minus irregularibus, ad orificium minutis hexagonis c. 15-20,« longis et c. 10-15/^ latis; stomatibus nuUis. L\rodomii dentés lineari-subulati, basi paulum connati, c. 0.32 mm. longi et basi c. 42 /^ lati, doro basi fuscescenti dein lutescenti laeves, linea media saepe anguste et longo perforata, ventro humile c. 15-18-lamellosi ; endostomium ? Sporac c. 20-25/^ magnae, fuscae, paulum scabrae. Calyptra cucuUata, c. 1 mm. longa, lutes- centi-viridescens apice fusca, parce (c. 6) pilosa, pilis lutescentibus erectis c. 0.4-0.5 mm. longis. Hondo : Prov. Idumo, Kiyomidudera (Leg. Masayoshi Nakaji ! 21. IX. 1913. et 11. I. 1914.). Species N. Immili Mitï. proxima, sed statura robustiore, foliis sicca haud rugulosis oculo nudo jam raptim cognoscenda. Nomen speciei ab Prov. Idumo. Wecl^era l^amaMirana Sh. Okamura. sp. nov. Habitat in truncis arborum. Piantae robustiusculae, caespito- sae, caespitibus superne laete viridibus inferne lurido-viridibus, moUiusculis, nitidiusculis, late extentis. Caiilis primarins elonga- tus, repens, divisus, hie illic fasciculatim fusco-radiculosus ; se- cundarius deflexus vel dependens, strictus vel fere strictus, c. 1-5 cm. longus, apice obtusus, simplex vel laxiuscule pinnatim ramosus. CONTKlBUTIONrS NO^AE AD FLOIÎAM BEYOPHYTON JAPONICAM. 43 sectione oblongiis, c. 0.24-0.32 x 0.16-0.24 mm. magnus, fasciculo centrali nullo, reti central! hyalino, cellnlis hexagonis c. 15-20/^ magnis, parietibus paulum incrassiusculis, cellulis periphericis c. 4-5-seriatis minoribus c. 1)-10,« magnis incrassatis luteis vel fuscis ; ramis c. 1.0-1.5 cm. longis, patentibus vel horizontalis, strictis, apice obtusis, simplicibus vel parce (1-3) ramulosis ; ramnlis obtusis ; caulibus secundariis ramis ramulisque dense Miosis com- planatis, cnm foliis c. 3-4 mm. latis, band flagelliformibus. Para- X20 X20 J^h/. lt>. Neckera kamakiirana. A. Folia (x20); B. Bracteae perichaetii intimae ( x 20) ; C. Sporangia ( x 10) ; D. Calyptra { x 20). phyllia nulla. Folia sicca immntata, superne paulum transverse rugulosa, dorsalia imbricata symmetrica, lateralia erecto-patentia vel divaricate -patentia, asymmetrica, breviter decurrentia, ovato- vel late ovate -oblonga, apice breviter et auguste vel latiore acuta, c. 2.1-2.4 mm. longa et 0.96-1.2 mm. lata, concava, baud plicata, marginibus e basi ad ultra medium uno latere auguste inflexo, in- fimis autem utroque auguste recurvis, superne minute serrulatis vel integerrimis ; nervo tenui, ultra medium evanido, in sectione trans - versali piano -convexo, dorso paulum prominenti, basi c. 40-45 fj lato et c. 15-20 /■« crasso, e 3-stratis cellularum subincrassatarum 44 ART. 4. SHÜTAI OKAMÜEA : uniformiuin coniposito ; cellulis chlorophyllosis, laevibus, parietibus incrassatis, superioribus rhombeis vel rhomboideis, c. 15-30 /^ longis et c. 7-8 /^- latis, in medio folii broviter linearibas c. 25-40 // longis et c. 5-7 p- latis, basilaribus c. 40-60 fj- longis et c. 7-8 /^ latis, in- fimis brevioribus c. 30-40 ,« longis et c. 9-10 /« latis inter sese porosis saepe Inteis, angularibus infimis rectangulis c. 15-20 <« longis et c. 9 ,« latis, angularibns supernis minntis quadratis c. 7- 9 [J- magnis luteis. Inßorescenüa autoica ; flores iitriusque sexus geminif ormes, in caulibus secundariis vel ramis ; folia perigonialia intima obovato-oblonga, apice subito in acumen breve subulato- acutum contracta, c. 0.75 mm. longa et c. 0.36 mm. lata, enervia ; antheridia c. 7, c. 0.4 mm. longa, paraphysibus paucis hyalinis. Ramulus perlchaetialis hand radiculosus. Bracteac pcridmetii in- timae basi alte vaginatae et obovato-oblongae, subito in acumen breviter et latiore subcuspidatum constrictae, c. 2.5-3.0 mm. longae et c. 1.0-1.1 mm. latae, margine ad apicem crenulatae vel integerrimae ; nervo tenui, ultra medium evanido ; cellulis su- perioribus late rhombeis vel rhomboideis, c. 15-20/^ longis et c. 7-9 A' latis, ceteroquin linearibus c. 30-100 A' longis et c. 7 /^ latis. Vaginida oblongo-cylindrica, c. 0.48-0.60 mm. alta et c. 0.24 mm. crassa, fusca ; paraphysibus numerosis, hyalinis. Seta c. 0.4-0.5 mm. longa et c. 0.12-0.16 mm. crassa, erecta, fusca, laevis. Theca immersa ovato-oblonga, symmetrica, erecta, c. 1.2-1.4 mm. longa et c. 0.64-0.75 mm. crassa, fusca, laevis ; cellulis exothecii rectan- gulis c. 30-56/^ longis et c. 15-20,« latis, ad orificium c. 3-4- seriebus minutis plane hexagonis ; stomatibus nuUis. Exostomli dentés basi connati, lineari-lanceolati, c. 0.35-0.4 mm. longi et basi c. 42 /^ lati, Strato dorsali laevi, inferne lutescenti-fuscescenti apice hyaline, linea media fere stricta et saepe hie illic paulum perforata, Strato ventrali hyaline laxe (c. 20-25) et humile lamelloso ; endo- CONTKTBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 40 stomium hyaliniim papillosum vel fere laeve ; corona basilaris haud producta; processus ignoti. Sporae c. 15-20;« magnae, virides, laeves. Calyptra cucullata, c. 1.2 mm. longa, lutea apice fusca, densiuscule pilosa, pilis lutescentibus, erectis, c. 0.4-0.5 mm. longis. Hondo : Prov. Sagami, Slmbusan [^ n^ tU], Kamaknm (Leg. Ktyotaka HisAucHi ! 5. YL 1915.). Species N. hum ill Mitt, affinis, sed bracteis perichaetii intimis obovato-oblongis, apice subito in acumen breviter subcuspidatum constrictis, calyptra densiuscule pilosis jam abunde diversa. Nomen speciei ab loc. Kamakura. Pinnatei'a forinosana Sh. Okamura. sp. noA'. Habitat in truncis arborurn. Plantac. robustiusculae, caespito- sae, caespitibus luteo-viridescentibus, haud nitidis, molliusculis. Calais Primarius elongatus, c. 5-8 cm. longus, densiuscule fascicu- latim fusco-radiculosus ; secundarius densus, strictus sed in statu sicco saepe curvatus, c. 1.5-5.0 cm. longus, inferne simplex stipiti- formis (stipite brevi c. 1-1.5 mm. vel longiore usque ad 10 mm. longo), dein dense regulariter pinnatim ramosus, in lineamentio oblongus vel oblongo-lanceolatus, cum ramis c. 7-20 mm. latus, superne saepe flagelliferous, sectione oblongus, c. 0.32-0.40 x 0.24-0.32 mm. mag- nus, fasciculo centrali nullo, reti centrali hyalino, cellulis hexagonis vel oblongo- hexagonis c. 15-20/-« magnis, parietibus incrassiusculis, cellulis periphericis c. 4~6-seriatis paulum minoribus valde incras- satis luteis vel fuscis ; ramis c. 7-13 mm. longis, erecto-patentibus, strictis, apice obtusis vel acutis, simplicibus vel irregulariter vel regulariter pinnatim ramulosis, saepe flagelliferois ; ramulis c. 5-7 mm. longis, obtusis vel acutis ; caulibus secundariis ramis ramulisque dense foliosis teretibus. ParaphylUa nulla. Folia sicca laxe ap- proximata longitudinariter tri-plicata, madida erecto-patentia ; folia 46 ART. 4. — SHUTAI OKAMUEA caulina e basi decurrentia et cordato-ovata lanceolato-lingulata, apice acuta vel breviter acuminata, c. l.G-2.0 mm. longa et basi c. 0.9- 0.95 mm. lata, concava, basi saepe leviter bi-plicata, marginibus e basi ad medium planis, integerrimis crenulatis vel apice minute serrulatis, haud limbatis ; nervo valido, infra summum apicem folii evanido, luteo, in sectione transversali plano-convexo, dorso promi- nent!, basi c. 70-80 />« lato et c. 40/^ crasso, e 3-4-stratis cellularum valde incrassatarum uniformium composite ; cellulis laevissimis, omnino incrassatiusculis, luniini- bus in medio folii 4-0-angulose rotundatis ovalis vel elipticis c. 9-20 [J- longis et c. 7 ,« latis, superioribus rliombeis vel rhom- boideis c. 20-30 /^ longis, basi- laribus breviter linearibus c. 15-30/^ longis et c. 4-6/^ latis, infimis luteis parce porosis, alari- bus indistinctis, marginalibus e basi ad medium c. 4-7-seriatis minoribus plerumque quadratis c. 6-7 i^ magnis ; folia ramea minora, c. 1.2-1.3 mm. longa et c. 0.6-0.64 mm. lata, ceteroquin ut folia caulina ; folia ramulina c. 0.8-0.9 mm. longa et c. 0.4-0.5 mm. lata, ceteroquin ut folia caulina. Caetera ignota. Formosa : Prov. Siuchiku-chü, Seuseki \ßi s] (Leg. Hisahiko Sasa- OKA ! 1. XII. 1912.). Species P. Mahinoi Beoth. similis, sed e cellulis superioribus et basilaribus foliorum fasiliter dignoscenda. Nomen speciei ab loc. Formosa. A.X20 Fig. -O. Piiuiatella forniosaua. A. FoHa (x20). CONTKIBUTIONES NOVAE AD FLORAM BllYOPHYTON JAPONICAM. 47 Tlianiniuni Fatrrfei Broth, et Par. Formosa : Prov. Sliincliiku-ch«'), Senseki (Leg. Hisahiko Sasaoka ! 1. XII. 1912.). Species nova ad flora m formosanam. LEMBOPHYLLACEAE. Dolichoiiiiti'a cyinbifolia (Lindb.) Broth, var. siihintegerrinia Sh. Okamuea. var. nov. Habitat in saxis. Plantae valde robustae, caespitosae, caepiti- bus viridibus vel liitescenti-viridibus seniora lutescenti-fuscis vel fnliginois, densiusculis, rigidiuscnlis. Caulis primarius repens, e- longatus, fusco-radicnlosus ; caulis secundarius arboriformis, erectus, c. 5-8 cm. altus, inferne stipitatus, superne dendroideam ramosus, stipite c. 1-3 cm. longo, e stipite flagella dependentia parvifolia fasciculatim fusco- vel rubiginoso-radiculosa emittentes, sectione teres, c. 0.4 mm. diam., fasciculo centrali nnllo, reti centi^ali hyalino, cellulis hexagonis c. 20 /^ magnis tenellis, cellulis periphericis c. 5-6-seriatis minoribus valde incrassatis fuscis ; ramis c. 3-4 cm. longis, curvatis, ramulosis ; ramulis c. 1-2 cm. longis, carvatis ; ramis ramulisque dense et tnmide foliosis, teretibus, apice obtusis. Folia sicca vix mutata, madida tumide imbricata vel patula, e, basi angusta elliptica vel subrotundato-elliptica, apice rotundato-obtusa vel brevissime apiculata, c. 1.2-1.8 mm. longa et c. 0.9-1.0 mm. lata, maxime cymbiformi-concava, apiculo hand recur vo potius leviter vel fortiter incurvo, marginibus basi recurvis, superne fere integris minute serrulatis vel parce minute duplicate -serrulatis ; nervo valide, infra apicem evanido, saepe furcato vel 1-2-nervu- loso ; cellulis laevissimis, anguste rhomboideis vel rhomboideo-linea- ribus, c. 28-40 /^ longis et c. 4-7 /^ latis, superioribus brevioribus 48 ART. 4. SHUTAI OKAMUEA c. 14-20/« longis, basilaribus linearibus c. 40-70/^ lungis et c. 5-7 ," latis inter sese vix porosis, alaribus paucis qiiadratis vel rectangnlis c. 14 /-« magnis fuscis. Caetera ignota. Hondo : Prov. Ise, in monte Gozaisliodake, Mie-gun (Leg. Kichitaro Murata! 15. Vin. 1914.). A typo foliis apice haud recurvis potius incurvis, marginibus superne fere integerrimis vel minute et parce serrulatis dignoscenda. Nomen varietatis ab foliis subintegerrimis. ENTODONTACEAE. Entodon avenosus Sh. Okamura. sp. nov. Habitat in arena fossae. Plantae robustiusculae, caespitosae, caespitibus viridibus, nitidiusculis, mollibns, late extentis. CauUs elongatus, prostratus, usque ad 12eni. longus, hie illic fasciculatim fusco-radiculosus, densiuscule pinnatim ramosus, sectione teres, c. 0.25 mm. diam., fascicule centrali arto c. 20 a« magno pauci-cellulari (c. 6), reti intermedio hyaline, cellulis hexagonis tenellis c. 15-20/^ magnis, cellulis periphericis c. 3-4-seriatis minoribus c. 10/^ magnis parce incrassatis chlorophyllosis sed postea fuscis ; ramis prostratis, e- longatis, usque ad 5 cm. longis, hie illic fasciculatim fusco-radicu- losis, simplicibus vel parce pinnatim ramulosis ; caulibus ramis ra- mulisque laxe foliosis, valde cora- planatis, obtusis, cum foliis c. 2.5 mm. latis. Parapliyllia nulla. Folia oc- tosticha, uniforma, symmetrica, sicca paulum inutata, madida folia dorsalia et ventralia laxe adpressa, lateralia patentia, valde concava, vix decur- rentia, e basi angusta late ovalio-elliptica, apice late et brevissime Flg. 21. Entodon arenosus. A. Folia (x20). CONTRIBUTIONES NOA"AE AD FLORAM BRYOPHYTON JAPONICAM. 49 cuspidata, c. 1.1-1.2 mm. longa et c. 0.8 mm. lata, marginibus basi reeurvis, saepe e medio ad basin cuspidis uno latere recurvo, superne serrnlatis, alls hand concavis ; nervis binis, tenellis, brevibus, c. 0.3-0.4 mm. longis ; cellulis valde clilorophyllosis, leavissimis ; linearibns, c. 70-110/^ longis et c. 7/^ latis, snperioribns brevioribus rhomboideis c. 20;" longis, basilaribns brevioribns inter sese hand porosis, alaribus numerosis quadratis c. 20// magnis vel breviter rectangulis c. 25-30 // longis et c. 20 /^ latis clilorophyllosis. Caetera ignota. Shikoku : Prov. lyo, Ugmsudani, Tamatsu-mnra, Nii-gun (Leg. Tsune- TARO Ota ! 19. XH. 1914.). Species E. Andoi Sh. Okamuea. valde affinis, sed foliis late ovali-eUipticis basi latioribus, fascicule centrali canlis pauci-cellulari dignoscenda. Nomen speciei ab vita in arena fossae. JEntodon dolichocuciillatus Sh. Okamura. sp. nov. Habitat in trnncis arborum. Plantae robustiusculae, dense de- presso-caespitosae, Intescenti-virides, nitidiusculae. Caulis prostra- tus, elongatus, c. 5-8 cm. longus, hie illic fasciculatim fusco-radicu- losus, dense et irregiilariter pinnatim ramosus, sectione plerumqne oblongns, c. 0.4-0.5x0.28-0.32 mm. magnus, fascicule centrali nuUo, reti centrali hyaline, cellulis hexagonis vel oblongo-hexagonis c. 15- 28 // magnis tenellis, cellulis periphericis c. 3-4-seriatis minoribus c. 7-15 // magnis hexagonis incrassatis lutescentibus ; ramis prostratis, c. 1.0-2.5 cm. longis, strictis, apice aliquando leviter curvatis, simplici- bus rarius parce vel pinnatim ramulosis, obtusis acutis vel attenuatis ; ramis ramuHsque dense foHosis, valde complanatis, cum fohis c. 2.5-3.0 mm. latis. Folia sicca vix mutata, dorsalia laxe imbricata, lateralia erecto-patentia, valde concava, e basi angusta ovato- vel 50 AET. 4. — SHUTAI OKAMURA oblongo-lanceolata, apice late vel anguste acuta, c. 1.9-2.1 mm. longa, basi c. 0.64-0.8 mm. et infime c. 0.48-0.56 mm. lata, mar- ginibus basi leviter recurvis, apice minute serrulatis ; nervis binis, c. 0.32-0.4-0.5 mm. longis, in sectione transversali fere planis, c. Id j^ crassis, e 2-stratis cellularum incrassatarum uniformium com- positis ; cellulis laevissimis, anguste linearibus, c. 90-140 /^ longis et c. 6-8 /J- latis, superioribus c. 50-70 /^ longis, basilari- bus infimis laxioribus c. SO- SO /^ longis et c. 13-16/^ latis parce porosis, alaribus numerosis quadratis sub- cpadratis vel breviter rectan- gulis c. 20-40 X 15-20 ft mag- nis hyalinis. lafiorescentia autoica ; flores utriusque sexus geminiformes in caule vel ramis ; folia perigonialia intima obovata, apice subito attenuata, c. 0.64-0.8 mm. longa et c. 0.32 mm. lata, concava, enervia ; anthe- ridia c. 6, c. 0.32 mm. longa, paraphysibus paucis, c. 0.32 mm. longis hyalinis. Ramulus perichactialis baud radiculosus. Bracteae jjerichaetii intimae basi semivaginatae lanceolatae apice sensim longe et anguste attenuatae, erectae, c. 2.1-2.4 mm. longae et c. 0.56-0.64 mm. latae, margine e basi ultra medium late recurvae, énerves vel indistincte binerves ; cellulis foliis similis. Vaginula cylindrica, c. 1.2-1.4 nnn. longa et c. 0.3 mm. crassa, pallida apice fusca ; paraphysibus paucis, hyalinis. Seta solitaria, c. 8-9 rarius 11 mm. longa, erecta vel leviter flexuosula, lutescens, laevis, sicca leviter torta. Thee a erecta, e collo obconico oblonge -cylindrica, symmetrica vel paulum asymmetrica, c. 1.4-1.6 mm. longa (collo FifJ. 22. Eiitodon dolichocucullata. A. Folia ( X 20) ; B. Theca ( x 10) ; C. Calyiitra ( x 10). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JArONICA:\r. 51 c. 0.4 mm. longo) et c. 0.56-0.6-0.8 mm. crassa, sicca deoperculata fere cylindrica et sub ore rarius paiilum contracta, rubiginosa, laevis ; celliilis exothecii quadrate -hexagonis oblongo-hexagonis rec- tangulisque, c. "20-70 X 20-35 /-« magnis, ad orificmm in seriebus 5 minoribus quadra tis hexagonis vel oblongo-hexagonis et transverse dilatatis ; stomatibus in collo paucis. Annubis nullus. Exostomil dentés subulato-lanceolati, basi connati, c. 0.32 mm. longi et basi c. 60-70,« lati, dorso basi (c. 70/^ longi) dense transverse saepe oblique vel longitudinaliter striatuli, dein dense papillosi, apice longitudinaliter striatuli, basi fusci dein luteo-fusci, saepe in linea media anguste perforati, ventre basi densiuscule dein remote et humile c. 10-lamellosi, hyalini, grosse papillosi ; endostomium luteo- fuscum ; corona basilaris paulum (c. 40 /^ alta) producta, papulosa ; processus dentium fere longitudinis, lineares, carinati, papillosi, rarius basi longitudinaliter striatuli. Sporae c. 9-10/^- rarius 14/^ magnae ocliraceae, minutissime papillosae. Opercunim e basi anguste conica rostratum, c. 0.8-0.9 mm. longum et c. 0.48-0.56 mm. diam., rubiginosum, rostro crasso stricto vel paulum curvato apice obtuse. Calyptra cucullata, c. 3.4-4.0 mm. longa, infra tliecam dépendons ibidemque apicem setae (usque ad c. 1.2-1.6 mm. longa) amplectens, pallida npice fuscescens, laevissima. Formosa : Prov. Tailiokn-chô, Sôzanslio [t^ Hi Œ] (Leg. Yaichi Shiivia- DA ! 1. I. 1914.). Species E. Drummondll (Be. euk.) Jaeg. subsimilis, sed foliis latioribus, apice anguste acutis, seta breviore, calyptra valde longissima, nee non peristomii structua exquo jam abunde diversa. Nomen speciei ab calyptra longissima. Pijlaisia laeto-vh'idis Sh. Okamura. sp. nov. Habitat in truncis. Plantae glaciles, dense depresso-caespito- 52 ART. 4. — SHCTAI OKAMURA sae, caespitibus late extentis, laete viridibus, serieeis, moUiuseulis. Caulis repens, usque ad 4 cm. longus, hie illic dense fasciculatim fasco-radiculosus, dense et regulariter pinnatini. ramosus, sectione teres, c. 0.2-0.24 mm. diam., fasciculo centrali valde arto pauci- cellulari, reti intermedio hyaline, cellulis hexagonis vel oblongo- hexagonis c. 15-20/^ magnis tenellis, cellulis periphericis c. 3-seriatis minoribus c. 7-15/^ lutescenti-fuscescentibus incrasatis ; ramis pro- stratis, c. 5-7 mm. rarius c. 10 mm. longis, strictis vel curvatulis, simplicibus rarius parce (1-3) ramulosis, dense foliosis subcom- planatis. Folia sicca laxe adpressa, plerumque homomalla, madida patentia, saepe homomalla, haud decurrentia, e basi angusta late vel anguste ovato-oblonga, apice acuto-attenuata vel sensim atte- nuata, c. 1.3-1.5 mm. longa et c. 0.48-0.64 mm. lata, valde con- cava, marginibus e basi ad basin acuminis leviter recurvis, inte- gerrimis rarius apice indistincte crenulatis ; nervis binis, tenellis, c. 0.24-0.4 mm. longis ; cellulis valde clüorophyllosis, laevissimis, pro- senchymaticis, in medio folii c. 40-50 ^ longis et c. 4-5 /^ latis, superioribus c. 28-40,« longis et e. 4-5/^ latis, basilaribus c. 50-70 // longis et c. 4-7 ,« latis, infimis brevioribus c. 25-40// longis, alaribus numerosis quadratis c. 15-20 [^ magnis valde chlorophyllosis. Inßorescentla autoica ; flores feminei in caule et ramis ; flores masculi in caule ; foha perigonialia intima ovata, apice acuta, c. 0.48 mm. longa et c. 0.24 mm. lata, enervia ; antheridia c. 5, c. 0.2 mm. longa, paraphysibus numerosis, hyalinis, c. 0.24-0.32 mm. longis. Ramulus perichaetialis parce fusco-radiculosus. Bracteae jjerichaetu intimae basi semivaginatae et lanceolatae apice sensim attenuatae, erectae. A. X20 Fig. 23. Pylaisia laeto-viridis. A. FoKa (x20); B. Tbeca (xlO). CONTRIBUTIONES NOVAE AD KLORAJt BRYOrHYTON JArONICA^M. 53 c. 2.0-2.4 mm. longae et e. 0.4-0. 5G imn. latao, integerrimae ; nervis binis indistinctis vel ntillis ; celhilis nt in foliis, sed longiori- bus angustioribus, basilaribus laxioribus anguste rectangulis c. 15-18/^ latis. Vaginula cylindrica, c. 1mm. longa et c. 0.3-0.4 mm, crassa, lutescenti-fuscescens ; paraphysibus numerosis hyalinis. Seta c. 6-7 mm. longa, leviter cnrvata, rnbra, laevissima, sicca torta. Theca erecta, e coUo distincto oblongo- cylindrica vel cylin- drica, symmetrica, c. 1.6-2.0 mm. longa (collo c. 0.4 mm. longo.) et c. 0.7-0.9 nnn. crassa, fusca, laevissima ; cellulis exothecii quadratis vel rectangulis saepe irregulariformis, c. 30-56 /^ longis et c. 20-40 [i- latis, ad orificium in seriebus c. 4 minoribns hexagonis transverse dilatatis ; stomatibns in coUo numerosis. An nidus triplex, c. 56/^ altus, basi fuscescens apice hyalinus, deciduus. Exostomu dentés lineari-lanceolati, basi connati, c. 0.3 mm. longi et basi c. 50 [J- lati, lutescenti vel fere hyalini, dense papillosi, saepe subopaci, intus laeves c. 20 humile lamellosi ; endostomium hyalinum, dense et grosse papillosum, hand opacum ; corona basilaris c. 70/^ alta, fere plana ; processus dentium longitudinis, lineari-subulati, divisi, ad tertiam partem soKtari dein ad apicem dentibus adhaerentes, hyalini, grosse papillosi; cilia nulla. Sporae c. 20-30,« magnae, virides, postea ferruginea, dense papillosae. Operculum e basi conica brevissime rostratum, c. 0.6 mm. longum et basi c. 0.56 mm. diam., rostro apice obtuso. Calyptra cucullata, c. 1.6-1.7 mm. longa, pallida, laevis. Hondo : Prov. Sagami, Kakuenjl in Kamakura (Leg. Kiyotaka Hisa- ucm! 5. VI. 1915.). Species a congeneribus exostomii dentés papillosi, processus divisi, ad tertiam partem solitari dein ad apicem dentibus adhae- rentes faciliter dignoscenda. Nomen speciei ab colore ea. 54 AET. 4. — SHÜTAI OKAMUEA : FABRONIACEAE. Schivetschhea 3Iatsiiviiirae Besch. Formosa : Taihoku (Leg. Yaichi Shimada ! 2. VI. 1914.) ; Prov. Shin- chiku-cliü, Chikunan-ippo [tj- M — %1 (Leg. Hisahiko Sasaoka ! 13. VI. 1912.). Species nova ad floram formosanam. HOOKERIACEAE. Mookeria nipponensis (Besch.) Beoth. Kj'iisliù : Prov. Ösumi, in insula Yaku-shima (Leg. Takanoei Iwaki ! 23. IX. 1914.). RHACOPILACEAE. Mhacopilutn aristatuin Mitt. Formosa : Prov. Taihokn-cliö, Urai (Leg. JuKiCHi Shieaga ! VII. 1911.). LESKEAOEAE. Haplohynienitifn bifomie Beoth. Kyüshü : Prov. Ösumi, in insula Yaku-shima (Leg. Takanoei Iwaki ! 23. IX. 1914). JELevpetineuron attenuates Sh. Okaäiuea. sp. nov. Habitat in truncis vel saxis. Plantae graciles, caespitosae, caespitibus densiusculis, lutescenti- vel fuscescenti-viridibus, haud nitidis. Caulis primarius elongatus, repens, stoloniformis, divisas, hic illic fasciculatim fusco-radiculosus ; secundariiis c. 7-12 mm. CONTRIBÜTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 55 longns, erectns siccitate apice circinatim involutus vel arcuatus, hnmidis apice pleriimque arcuatus, dense foliosus teres, apice ob- tusus vel breviter acutus saepe longiuscule (c. 5-7 mm.) flagellifor- mis, simplex vel parce (1-3) ramosus, sectione teres vel ellipticus, c. 0.24-0.3x0.16-0.20 mm. magnus, fasciculo centrali arto, reti intermedio hyaline, cellulis liexagonis c. 14-20/^ magnis tenellis, cellulis peripherisis c. 5-7-seriatis minoribus c. 5-10/^ magnis valde incrassatis lutescentibus vel lu- teo-fuscis ; ramis c. 2-5 mm. rarius c. 8 mm. longis, apice obtusis vel breviter acutis, plerumque arcuatis. Folia sicca adpressa, madida erecto-patentia, breviter decurrentia, oblonge - lanceolata, apice longe attenuata vel longiore acuta, c. 1.2-2.0 mm. longa et c. 0.4-0.6 mm. lata, basi bi-plicata, marginibus planis, apice serratis ; nervo crasso, lutescenti vel luteo-fusco, superne flexuoso, infra summum apicem folii evanido, in sectione transversal! piano- con vexo, dor so valde prominenti, basi c. 84/-« lati et c. 50 /^ crasso, e c. 5-6-stratis cellularum uniformium valde incrassatarum composito, cellulis ventralibus c. 8, dorsalibus c. 15-17 minoribus ; cellulis laminalibus valde chlorophyllosis, in Seriebus obliquis dispositis, subquadratis vel subrhombeis, c. 5-7/^ magnis, parietibus incrassiusculis, luminibus quadratis ellipticis vel oblongis, omnino leavibus et pellucidis. Caetera ignota. Fig. 24. Herpetineurozi attenuatus. A. Planta, a. in statu sicco ( x 1) ; B. FoKa (x30). 56 ART. 4. — SHÜTAI OKAilURA; Hondo : Prov. Sagami, Kamaknra (Leg. Kiyotaka BüsAUCm ! 1. TTT. 1914.). Species distinctisma, a congeneribus foliorum forma faciliter dignoscenda. Nomen speciei ab foliorum formis. Leskea pusilla Mitt. in Trans. Linn. Soc. London. 2nd. Ser. Bot. Vol. m. part 3. p. 188 (1891); Icônes Plant. Koisikavenses. Vol. m. No. 1. p. 17. Taf. 154. (1915). Habitat in truncis arborum defectrum. Plantae tenellae, cae- spitosae, caespitibus fuscescenti-viridibus, hand nitidis, mollinsculis, late extentis. Caulis repens, usque ad 3 cm. longus, hie illic fas- ciculatim fusco-radiculosus, dense et subirregulariter pinnatim ra- mosus, sectione teres, c. 0.12-0.16 mm. diam., fascicule centrali nullo, reti centrali hyaline, cellulis hexa- gonis c. 14/^ magnis, cellulis periphericis c. 2-3-seriatis paulum minoribus incrassatis lutescentibus ; ramis prostratis vel as- cendentibus, simplicibus rarius 1-2- ramulosis, strictis vel curvatulis, apice obtusis vel acutis, usque ad 6 mm. longis, sicca arcuatis vel flexuosis, plerumque apice hamulosis, densiuscule foliosis sub- complanatis. ParaphylUa nulla. Folia sicca adpressa vix crispata, madid a erecto- patentia, vix decurrentia, valde concava, stricta vel curvatula, e basi contracta ovato-lanceolata vel oblongo-lanceolata, apice breviter attenuata saepe leviter incurva vel recur va, summo apice acuta vel subobtusa, c. 0.8-1.2 mm. longa et c. 0.3-0.4 mm. lata, marginibus basi leviter recurvis, apice suberectis, integerrimis Fig. 25. Leskea pusilla. A. Folia ( X 30) ; B. Tlieca (xlO). CüNTRIBUTIOXES NOYAE AD FLOKAM BRYOPHYTON JAPONIC AM. 57 vel apice crenulatis ; nervo valido, infra summum apicem folii evanido, e medio ad apicem leviter flexuoso, lutescenti, in sectione transversali piano- con vexo, dorso valde prominenti, basi c. 42 /^ lato et c. 28 /^ crasso, e 3-4-stratis cellularum uniformium in- crassatarum composite, ceirulis dorsalibus c. 8, ventralibus c. 4 ; cellulis laminalibus in medio folii quadratis subquadratis ovali- quadratis vel rectangulis, c. 7-9-14/^ longis et c. 7-9/^ latis, superioribus ovalis vel oblongis, basilaribus quadratis vel rectangulis c. 7-10-18/^ longis et c. 7-10// latis, angularibus quadratis, omnino laevibus vel parce humile papulosis baud opacis. Inflorescentia autoica ; flores masculi in caule vel ramis geminiformes ; folia perigonialia 5-8, intima quadrato-obovata, apice subito acuto- attenuata vel cuspidata, c. 0.32 mm. longa et c. 0.24 mm. lata, concava, enervia ; antberidia 5-8, parapliysibus perpaucis, hyalinis vel fuscescentibus ; flores feminei in caule. Ramulus pericliaetialis valde radiculosus. Bracteae perichaetü intimae basi semivaginatae oblongo-lanceolatae apice subito lanceolate -subulate attenuatae, c. 1.6-2.0 mm. longae, integerrimae ; nervo valido, infra summuna apicem evanido, lutescenti-fusco ; cellulis prosenchymaticis, basilari- bus rectanglaribus laxioribus, laevibus. Vaglnula cylindrica, c. 0.9-1.0 mm. alta et c. 0.25-0.30 mm. crassa, fuscescens. Seta c. 10-15 mm. longa, stricta curvatula vel flexuosula, sicca per totam longitudinem torta, inferne rubra, superne fuscescens, laevis. Theca symmetrica vel paulum asymmetrica, oblonge -cylindrica, c. 1.6-2.5 mm. longa et c. 0.4-0.56 mm. crassa, stricta vel curvatula, fusca, laevis, brevicoUis (coUo obconico c. 0.24 mm. longo.), sicca de- operculata sub ore paulum contracta ; cellulis exothecii rectangKs quadratis vel oblonge -quadratis, ad orificium in seriebus c. 5 minutis hexagom's ; stomatibus in collo sat numerosis, phaneroporis. Peri- stomium duplex ; exostomii dentés lineari-subulati, basi connati, c. 58 ART. 4.— SHÜTAI OKAMURA : 0.56 mm. longi et basi c. 0.07 mm. lati, strato dorsali luteo, basi (usque ad c. 0.14 mm. alto.) transversim striolato inde ad apicem minutissime papilloso, aliquando superne pertuso, linea media flexuosula, strato ventrali paulum latiore dense (c. 35.) et alte lameUoso ; endostomium luteum, minutissime papillosum ; corona basilaris usque ad 0.1 mm. alta, leviter plicata ; processus dentium longitudinis, anguste lineari-subulati, carinati in carina perforati, ventre c. 15-trabeculati ; cilia nulla. Sporae c. 15/>« magnae, fuscescenti-virides, minutissime papillosae. Caetera desiderantur. Hondo : Prov. Etcliü, Tsukioka-mura, Kamisliinkawa-gun (Leg. HiSA- HIKO Sasaoka ! 30. IX. 1913.). Prov. Miisaslii ; Öraori, Kasuga-jinslia (Leg. Shütai Okamura! 2. H. 1913.). Distr. endemica. Haxylocladiiiiti cajnllattini (Mitt.) Broth. Formosa : Prov. Sliinchiku-cliô, Cliiknuau-ippo (Leg. Hisahiko Sasa- oka ! 14. V. 1912.). Species nova ad floram formosanam. Haplocladium latifolium (Lac.) Broth. Formosa : Prov. Tailioku-chö, Taihoku (Leg. Hisahiko Sasaoka ! 31. X. 1912.). Tliuidiiim jaiionicum Doz. et Molk. Formosa : Prov. Shincliiku-cliö, Naroyama (Leg. Hisahiko Sasaoka ! 1. Xn. 1912.). Kyüslm : Prov. Ösumi, in iusiüa Yaku-shima (Leg. Taka- KORi Isvaki ! 21. IX. 1914.). CONTBIBUTIONES NOVAE AD ITLOKAM BRYOPHYTON JABONICAM. 59 HYPNACEAE. Drepanocladus uncinatus (Hedw.) Warnst. Korea : Prov. Kaukyo-Nandö, Chôdadô [ë tô M]^ Sansui-gun [h 7K M] (Leg. Takenoshin Nakai ! 24. Vil. 1914.) ; iu monte Roliô (Leg. Takenoshin Nakai! 11. Vn. J914.). Species nova ad floram koreanam. Hygrohypmun ochraceuiu (Trun.) Broth. Hondo : Prov. Iwashiro, Kamanumahara (Leg. Genichi Koidzumi ! VUE. 1913.). Species nova ad floram japonicam. JRhytidiadelphus calvescens (Wils.) Broth. Korea : iu monte Koliù (Leg. Takenoshin Nakai ! 11. YII. 1914.). Species nova ad floram koreanam. JRhytidiadeljyhus squarrosus (L.) Warnst. Korea : iu monte Koliô (Leg. Takenoshin Nakai ! 10. YII. 1914.). Species nova ad floram koreanam. Rhytidiadelphus triquetruvi (L.) Warnst. Korea : iu monte Rohô (Leg. Takenoshin Nakai ! 11. YII. 1914.). Species nova ad floram koreanam. ItJiytiditirii rugosiun (Ehrh.) Kindb. Korea : iu monte Kohö (Leg. Takenoshin Nakai ! 11. YII. 1914.) ; iu monte Hakutôsan [â M Ui] (Leg. Takenoshin Nakai ! 10. YHI. 1914.). Species nova ad floram koreanam. 60 ART. 4. — SHÛTAI OKAMURA : JTylocoiniuni jji/renniciini (Spkuc.) Ltndb. Korea : iu monte lloliö (Leg. Takenoshin Nakai ! 11. VU. 1914.). Species nova ad floram koreanani. Hyloconiiuni splendens (Hedw.) Schimp. Korea : in monte Rohö (Leg. Takenoshin Nakai ! 11. VIT. 1914 ). Species nova ad floram koreanam. Hypnuni Schreberi Wild. Sachalin : Sicca (Leg. Jüsuke Susuki ! 1. V. 1913). Korea : in monte Rohô (Leg. Takenoshin Nakai ! 10. VIT. 1914.). Species nova ad floram sachalinam et koreanam. JSctrojmtheciuni Shiragae Sh. (3kamura. sp. nov. Habitat in truncis arborum. Plantae graciles, caespitosae, caespitibus liitescenti-viridibus, nitidiusculis, molliusculis, densissimis, late depresso-extentis. Caulis elongatus, repens, c. 5-7 cm. longus, per totam longitudinem fasciculatim fusco-radiculosus, divisas, dense et regulariter pinnatim ramosus, sectione oblongus, c. 0.24-0.32 x 0.16-0.24 mm. magnus, fasciculo centrali arto panci-cellnlari, reti intermedio hyalino, cellulis hexagonis vel oblongo-hexagonis c. 15- 30 A« magnis tenellis, cellulis peripliericis c." 3-seriatis minoribus in- crassatis lutescentibus vel fuscis ; ramis c. 3-5-7 mm. longis, trans- verse patentibus et prostratis vel patulis et adscendentibus, strictis, simplicibus rarius parce vel regulariter pinnatim ramulosis apice ob- tusis, dense Miosis complanatis, cum foliis c. 1.2 mm. latis. Para- jdhyllia nulla. Folia quasidisticha, baud liomomalla, sicca laevia, ovato-lanceolata, apice subulato- acuminata, c. 0.9-1.2 mm. longa et c. 0.28-0.32 mm. lata, concava, baud plicata, marginibus basi levi- ter recurvis, inferne integerrimis, e medio ad apicem argute serra- COXTRIBÜTIOXES NOVAE AD FLOEAM BRYOPHYTON JATONICAM. 61 A.X30 FifJ, 2G. Ectropothecitini Shiragae. A. Folia (x30); B. Tbeca (xlO). tis ; nervis binis, c. 0.24-0.32 mm. longis, in sectione transversali ferc planis, baci c. 15-20/^ latis, e 2-stratis cellnlarum incrassata- rum nniformium compositis ; cellulis laevissimis, angnste linearibus, c. 30- 56 /-« longis et c. 3 /^ latis, marginalibiis non difformibus, basilaribus c. 4,« latis, infimis laxioribus rectangnlis c. 20-30 /-^ longis et c. 7-9 /^ latis hyalinis, alaribus 1 rarius 2 magnis oblongis c. 30-40 /^ longis et c. 15-20/^ latis subvesiculae- formibus hyalinis, siipra alaribus paucis quadratis vel breviter rectangnlis minori- bus c. 15/^ magnis hyalinis. InßorescenUa autoica ; flores utriusque sexus geminif ormes in canle ; folia perigonialia intima rotundato- ovata, apice subulate -attenuata, c. 0.5 mm. longa et c. 0.28 mm. lata, concava, enervia ; antheridia numerosa (c. 15), c. 0.2 mm. longa, paraphysibus paucis, c. 0.24 mm. longis, hyahnis. Ramulus jjerichaetialis parce radiculosus. Bradeae perichaetii intimae basi longe semivaginatae lanceolatae in acumen sensim elongatum pili- forme falcatum serratum attenuatae, c. 2 mm. longae et c. 0.4 mm. latae, hand plicatae, margine e basi ad ultra medium late recurvae, énerves vel obsolete binerves ; cellulis linearibus, c. 40-70 /^ longis et c. 5-6 /^ latis, basilaribus infimis laxioribus rectanguhs c. 30-45 i" longis et c. 10-15 />« latis. Vaginula cylindrica, c. 0.8 mm. alta et c. 0.3 mm. crassa, pallida ; paraphysibus numerosis hyahnis. Seta c. 8 mm. longa, tenuis, c. 0.08 mm. crassa, substricta, rubra, laevis, sicca paulum torta. Theca nutans, e coUo obconico distincte ovahs, symmetrica, c. 0.8-0.9 mm. longa (collo c. 0.10 mm. longo.) et c. 0.56 mm. crassa, fusca, superne paulum mamillosa, sicca deoper- culata sub ore valde contracta hand curvata, leavis ; cellulis exo- 62 ART. 4. — SHÜTAI OKAMUEA : thecii rotundato-quadratis paulum collenchymaticis, c. 20-35 // magnis, ad orificium in seriebns c. 4-5 minoribns hexagonis c. 10-15 />« magnis; stomatibus nullis. Anmihis duplex, c. 45// altus, basi fuscescens inde liyalinus, deciduns. Exostomii dentés subulato- lanceolati, basi connati, c. 0.32 mm. longi et basi o,. 10 f^ lati, dorso basi lutei transversim striatuli, superne hyalini grosse papillosi, ventro hyalini dense (c. 35) et alte lamellosi ; endostomium corona basilaris c. 140 /^ alta, ochracea, laevis ; processus dentium fere longitudinis, ochracei, papillosi, carinati in carina auguste per- forati ; cilia bina, bene evoluta, hyalina, papulosa, nodulosa. Spo- rae c. 7-8 /^ magnae, ochraceae, minntissime papillosae. Operculum a basi conica mamillosa rostratum, c. 0.5 mm. longum et basi e. 0.5 mm. diam., rostro brevi c. 0.2 mm. longo erecto laevi. CaJyptra cucullata, c. 1.5-1.6 mm. longa, pallida, laevis. Formosa : Prov. Taihoku-clK"), Urai (Leg, Jükichi Shiraga ! Vu. 1911.). Species E.planulo Card, proxima, sed foliis subulato-acumina- tis, cellulis angustioribus, seta breviore diversa. Nomen speciei in honoren Coli. Dom. Jukichi Shiraga. Stereodon cymbifolius Sh. Okamuea. sp. uov. Habitat in saxis. Plantac subgraciles, dense caespitosae, caes- pitibus late extentis, viridibus seniora fusco-viridibus, serieeis, mollioribus. Caulis elongatus, repens, usque ad 5 cm. longus, hie illic fasciculatim fusco-radiculosus, dense et subirregulariter pin- natim ramosus, sectione teres, c. 0.1ß-0.24mm. diam., fasciculo centrali nuUo, reti centrali luteo, cellulis hexagonis vel oblongo- hexagonis tenellis c. 20-25 // magnis, cellulis periphericis c. 2-3- seriatis minoribus c. 15// magnis incrassatis fuscis, externis luteis et membrana externa tenella ; ramis 8-15 mm. rarius usque ad 25 mm. longis, simplicibus vel irregulariter pinnatim ramulosis, adscendenti- COMTEIBUTIONES NOVAE AD FLORAM BEYOrHYTON JAPONICAM. 63 bus vel fere erectis aliquando prostratis, strictis vel curvatulis, apice obtusis sed sicca aciitis, dense foliosis complanatis, cum foliis 1.2-1.6 mm. latis. Paraphyllia nulla. Folia sicca laxe adpressa baud liomomalla nee falcata, madida erecto-patentia, e basi angus- ta late oblonga vel obovato-oblonga, subito in acumen saepe semitortum plus minus elongatum subulatum sensim attenuatum constricta, c. 0.8-1.2 mm. longa et c. 0.4- 0.56 mm. lata, cymbiforme concava, non plicata, marginibus basi leviter recurvis dein late incurvis, superne argute serratis ; ner- vis binis, tenellis, c. ('.16-0.3 mm. longis ; cellulis laevissimis, auguste prosenchymaticis, c. 80-120 u- longis et c. 4-6 /^ latis, superiori- bus brevioribus c. 40 /^ longis, basilaribus infimis abbreviatis et laxioribus c. 35-40 <« longis et 7-10 /^ latis aureis, alaribus 3 magnis oblongis vesiculae- formibus c. 30-50/^ longis et c. 20-35^ latis, lutescentibus vel fere hyalinis, supraalaribus perpaucis c. 4-5 quadratis vel rectangulis lutescentibus vel fere hyalinis. Inflorescentia dioica ; flores feminei in caule et ramis ; flores masculi in ramis geminiformes, minoriores, c. 0.2-0.24 mm. longi ; folia perigonialia intima ovata, apice acuta, concava, enervia, c. 0.18-0.24 mm. longa et c. 0.08-0.1 mm. lata; antheridia pauca (c. 2-3), c. 0.14 mm. longa, paraphysibus per- paucis, c. 0.1 mm. longis, hyalinis. Ramulus perichaetialis valde fusco-radiculosus. Bracteae perlchaetu intimae lanceolatae, apice sensim elongate attenuatae, erectae vel recurvae, c. 1.2-1.4 mm. longae et c. 0.3-0.4 mm. latae, concavae, leviter plicatae, superne argute serratae ; nervis binis, c. 0.4-0.5 mm. longis, saepe in- Fiff. 27. Stereodon cymbifoliiTS. A. Folia ( X 30) ; B. Sporangia (xlO). 64 ART. 4. — SHUTAI OKAMURA : distinctis ; cellulis linearibus, basilaribus laxioribus, fere rectangulis aureis. Vaginu/a cylindrica, c. 1.1-1.3 mm. longa et c. 0.3 mm. crassa, brumiea ; paraphysibus perpaucis. Seta c. 10-15 mm. longa, stricta vel leviter flexuosula, rubiginosa, laevis, sicca torta. Tlieca obliqua vel horizontalis, e coUo obconico anguste ovato- oblonga, deoperculata fere cylindrica, c. 1.6-1.8 mm. longa (collo c. 0.3 mm. longo.) et c. 0.8-0.85 mm. crassa, panlum asymmetrica, sicca deoperculata leviter curvata, sub ore paulum contracta, laevis, castanea ; cellulis exotliecii rectangulis, c. 40-56;^ longis et c. 20-30/^ latis, parietibus incrassatis c. 10-14/^ crassis, ad orificium in seriebus tribus minutis quadratis vel liexagonis c. 20 f^ magnis ; stomatibus in collo paucis. Annulus duplex, c. 40 /^ altus. rubiginosus, deciduus. Exostomü dentés subulato-lanceolati, basi connati, c. 0.44 mm. longi et basi c. 60-70/^ lati, dorso luteo- fusci, dense transversim striatuli, apice papillosi, liiiea media flexuosula, ventre dense (c. 35) et altissime lamellosi ; endostomium luteum, minute papillosum ; corona basilaris c. 0.15 mm. alta ; processus dentium longitudinis, carinati in carina anguste perforati ; cilia bina, bene evoluta, nodulosa, papiUosa. Sporae c. 20-25/^ magnae, virides, laevissimae. Operculum e basi conica longe rostratum, c. 0.9 mm. longum et basi c. 0.6 mm. diam., rostro stricto, c. 0.64 mm. longo. CaJyptra cucuUata, c. 1.4-1.6 mm. longa, basi lutea superne fuscescens, laevis. Hondo : Prov. Etcliû, in monte Kurobe-yama, Sliimosliinkawa-gun (Leg. HiSAHiKO Sasaoka ! 3. Yin. 1915.). Species S. Henoni (Duby.) Mitt, affinis, sed statura paulum graciliore, seta breviore praecipue foliorum forma longe diversa. Nomen speciei ab foliis cymbiforme concavis. CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 65 Stereodon fertilis (Sendtn.) Linde. Korea : in monte Gatokurei (Leg. Takenoshin Nakai ! 5. VII. 1914.) ; in monte Hakutosan (Leg. Takenoshin Nakai ! 10. VIII. 1914.). Species nova ad floram koreanani. Stereodon pliimaefomiis (Wils.) Mitt. Formosa : Prov. Tailiokuclio ; in monte Mentenzan [M ^ Uj] (Leg. Yaichi Shimada ! 31. X. 1914.) ; Tailioku (Leg. Hisahiko Sasaoka ! 2. XL 1912.). Isoxjterygiwui expallescens Levier. Formosa : Prov. Tliilioku-cho ; Urai (Leg. Jukichi Shiraga ! VII. 1911.). Species nova ad floram formosanam. Isopterygitim Giraldii (C. Müll.) Par. Formosa : Tailioku (Leg. Yaichi Shimada ! 20. VI. 1914.). Isopterygiuiu Misauchii Sh. Okamura. sp. nov. Habitat in truncis arborum defectrnm. Plantae graciles, cae- spitosae, caespitibus laete viridibus, late extentis, moUibns, nitidis. Calais repens, usque ad 4 cm. longus densiuscule rubiginoso-radi- culosus, laxe et irregulariter pinnatim ramosus, sectione teres, c. 0.16 mm. diam., fasciculo centrali nullo, reti centrali hyaline, cellulis tenellis hexagonis c. 15-20/^ magnis. ceUulis periphericis c. 2-3- seriatis minoribus c. 7-9 /^ magnis hand incrassatis lutescenti-viri- descentibus vel lutescenti-fuscis ; ramis prostratis, usque ad 1.5 cm. longis, simplicibus, densiuscule foliosis valde complanatis, apice ob- tusis vel breviter attenuatis. Folia sicca patentia leviter plicata, madida patentia non plicata, baud decurrentia, symmetrica, e basi ovata lanceolata, in acumen elongatum angustum subulatum sensim 66 ART. 4. — SHUÏAI OKAMURA attenuata, c. 1.9-2.4 mm. longa et c. 0.5-0.64 mm. lata, concava, marginibus basi erectis, ubique praecipue ad apicem argute serratis ; nervis binis, tenuis, brevibus c. 0.64 mm. longis, viridibus ; cellulis valde chlorophyllosis, laevissi- mis, linearibus, c. 00-90 /^ longis et c. 7-8 /^ latis, basilaribus infimis haud porosis rectangulis c. 25-35/^ longis et c. 10/^ latis chlorophyllosis • saepe fuscis vel hyalinis, alaribus inconspicuis. Inßorescentia autoica ; flores utriusque sexus in caule geminiformes ; folia perigonialia intima ovata apice sensim attenuata, c. 0.48-0.72 mm. longa et c. 0.30-0.36 mm. lata, concava, enervia, margine crenulata ; antheridia c. 8, paraphysi- A^x20 bus paucis brevibus. Ramulus ^Jer^c/^ae^^a//s Fig. 28. isoptery- yaldc fusco-radiculosus. Bracteae perichaetii in- gium Hisauchii. A Folia (X 20)- timae e basi semi-vaginante sinsim auguste B. Sporangia (X 10). subulatae, c. 1.7-2.0 mm. longae et c. 0.4-0.5 mm. latae, subula erecta vel recurva, marginibus e medio ad apicem argute serrulatis ; nervo nuUo ; cellulis linearibus, chloro- phyllosis, saepe apice dorso prominente papillosis, basilaribus rectangulis fuscis. Vaginula cylindrica, c. 0.8 mm. longa, lutescenti- viridescens apice fusca. Seta c. 1.0-1.5 cm. longa, tenuis, stricta, basi rubra superne lutescenti-rubescens, laevissima, sicca torta. Theca inchnata vel horizontalis, e collo distincto conico oblonga, asymmetrica, leviter gibbosa, sicca deoperculata sub ore valde contracta haud plicata, c. 1.4-1.9 mm. longa (collo c. 0.4-0.5 mm. longo.) et c. 0.6-0.8 mm. crassa, fusca, leavis ; cellulis exothecii quadratis vel rectangulis, c. 25-50 />« longis et c. 20-30/^ latis, ad orificium in seriebus 3-5 minutis subquadratis ; stomatibus in collo numerosis. Annulus duplex, c. 50-50/-« altus, e duplici série CONTRIBUTIONES NOYAE AD FLORAM J3RY0PHYT0N JAPONICAM. 67 cellularum, deciduus. Exostomii dentés basi connati, subulato- lanceolati, c. O.oö mm, longi et basi c. 70/^ lati, dorso lutoi, basi transversim striatuli, superne papillosi, ventro dense (c. 30) et alte lamellosi ; endostomium liyalinum, minutissime papillosum ; corona basilaris ad medium dentium producta ; processus dentium fere longitudinis, carinati in carina vix perforati ; cilia terna rarius bina, bene evoluta, nodulosa, breviter appendiculata, Sporae c. 9-12/^ magnae, luteo-virides, minutissime papillosae. Operculum e basi conica breviter rostratum, c. 0.56 mm. longum et basi c. 0.56 mm. diam., rostro crassiusculo recto obtuso. Cafyptra cucuUata, c. 2.8 mm. longa, inferne lutescenti-viridescens superne fuliginosa, laevis. Hondo : Prov. Sagarai ; Sagijama, Yokohama (Leg. Kiyotaka Hisa- UCHI ! 27. V. 1915.) Habitus statura faciesque omnino /. taxirameum (Mitt.) Jaeg., foliis autem apice anguste attenuatis, inflorescentia autoica, operculo breviore rostrato longe diversis. Nomen speciei in honoren Coll. Dom. Kiyotaka Hisauchi. Jsopterygiuni taxiratneiim (Mitt.) Jaeg. Formosa : Prov. BincLiku-cliô ; Cliikimau-ippo (Leg. Hisahiko Sasa- oivA ! 5. XH. 1912.). Vesicularia Sasaokae 8h. Okamuea. sp. nov. Habitat in rupibus. Plantae tenellae, dense deprrso-caespitosae, lutescenti-virides, molles, nitidiusculae. Caulis repens, c. 4-6 cm. longus, hie illic dense fasciculatim rubiginoso-radiculosus, denso et regulariter pinnatim ramosus, sectione teres vel ellipticus, c. 0.24- 0.32 mm. diam., fasciculo centrali nullo, reti centrali hyalino, cellulis tenellis hexagonis c. 15-28 />« magnis, cellulis periphericis c. 2-3-seriatis minoribus c. 10-15/^ magnis incrassiusculis luteis vel 68 ART. 4. — SHÜTAI OKAMÜRA. fiiscescentibus ; ramis plerumque 4-6 mm. rariiis usque ad 8 mm. longis, patulis vel horizontalis, strictis, obtusis, cum foliis c. 1.5-2.0 mm. latis, simplicibus aliquando parce (1-2) ramulosis ; caulibus ramisque dense foliosis complanatis. Folia sicca vix mutata apice flexuosula leviter recurva vel stricta, madida erecto-patentia, haud homo- malla, concava ; folia lateralia pau- lum asymmetrica, ovato-oblonga, sensiu in acumen elongatum filifor- me strictum vel leviter flexuosum attenuata, c. 1.5-1.8-2.0 mm. longa et c. 0.56-0.64 mm. lata, margini- bus in acumen minute serrulatis A. X20 f.d. Fig. 29. Vesicularia Sasaokae. A. Folia : f.d. foHa dorsalia ; f .1. folia lateraUa vcl intCgCrrimis ; folia dorSalia fcre (x20); B. Sporangia ( x 10). symmetrica, oblongo-lanceolata, apice breviter attenuata, c. 1.4-1.6 mm. longa et c. 0.5 mm. lata; nervis binis, brevibus, tenellis, c. 0.32-0.4 mm. longis ; cellulis chlorophyllosis, laevissimis, parietibus incrassiusculis, utriculo primordiali distincto, in medio folii late rhomboideis, c. 40-56 ft longis et c. 15-18,« latis, in acumen linearibus c. 110-140/^ longis et c. 7 ij- latis, basilaribus lineari- rhomboideis c. 80 /^ longis et c. 15-18,« latis, infimis fere rectangulis vel oblongis c. 40-50/^ longis et c. 20-28 /-« latis, alaribus indistinctis. Inßorescentia autoica ; flores utriusque sexus geminiformes in caule ; folia perigonialia intima rotundato-ovata, apice subito breviter attenuata, c. 0.48 mm. longa et c. 0.25 mm. lata, concava, enervia ; antheridia c. 10, c. 0.25 mm. longa, paraphysibus paucis hyalinis brevibus. Ramulus jyeiHchaetialis valde rubiginoso-radiculosus. Bracteae perlchaet'd in- timae oblongo-lanceolatae, in acumen sensim elongatum piliforme erectum vel leviter recurvum integerrimum vel minute serrulatum CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JArONICAM. 69 attenuatae, c. 2.2 mm. longae ct basi c. 0.64-0.7 mm. latae ; nervis binis, indistinctis, brevibiis, c. 0.4 mm. longis ; cellulis nt in foliis. Vagimda cylindrica, c. 0.8 mm. longa et c. 0.4 mm. crassa, fiisca ; paraphysibus paucis, hyalinis, brevibus, c. 0.5 mm. longis. Seta <3. 10-15 mm. longa, pauliim flexuosa, rubra apice lutescenti-rubra, laevis, sicca leviter torta. Tlieca pendula vel nutans, oblonga, c. 1.4-1.5 mm. longa et c. 0.9 mm. crassa, brevicollis, fusca, laevis, sicca deoperculata sub ore contracta ; cellulis exothecii quadratis rectangulis vel hexagonis, c. 25-56 /-« longis et c. 25-40 /^ latis, ad orificium in seriebus c. 3-4 minoribus quadratis vel hexagonis c. 15,« magnis ; stomatibus nullis. Annuliis duplex, c. 56/^ altus, hyalinus, deciduus. Exostomii dentés subulato-lanceolati, basi «onnati, c. 0.48 mm. longi et basi c. 84 /-« lati, dorso e basi ad medium luteo-fusci et dense transversim striatuli, dein hyalini laxe et grosse papillosi, linea media flexuosula, ventro dense (c. 40) et alte lamellosi ; endostomium luteum, minute papillosum ; corona basilaris c. 0.16 mm. alta ; processus dentium longitudinis, carinati in carina anguste perforati, grosse papillosi, apice hyalini ; cilia bina, bene evoluta, hyalina, papulosa. Sporae c. 7-9 /i magnae, fuscescentes, laeves. Operculum e basi convexo-conica acute api- oulatum, c. 0.7 mm. longum et c. 0.7 mm. diam. CaJyptra cucul- lata, c. 2 mm. longa, lutescens, laevis. Formosa : Prov. Taihoku-cliô, Eifukiisan [/tc Is Oj] (Leg. Hisahiko Sasa- OKA ! 24. n. 1912.). Species V. reticulata (Doz. et Molk.) Beoth. valde affinis. sed foliis longiore acuminatis faciliter dignoscenda. Nomen speciei in honoren Coll. Dom. Hisahiko Sasaoka. Vesicularia Shhnadae Sh. Okamura. sp. nov. Habitat in rupibus. Plantae robustiusculae, dense depresso- 70 ABT. 4. — SHÜTAI OKAMUß A : caespitosae, luteo-fiiscescentes, nitidiusculae, rigidiusculae. Caulls repens, elongatus, c. 4-7 cm. longus, per totam longitudinem faseica- latiin fusco-radiculosus, divisiis, dense et regiilariter pinnatirn ra- mosus, sectione teres vel oblongus, c. 0.40-0.48 x 0.25-0.32 mm. magmis, fasciculo contrali nullo, reti centrali hyalino, cellnlis tenellis hexagonis vel oblonge -hexagonis c. 25-40," magnis, cellulis periphericis c. 2-3-seriatis minoribns c. 14/^ magnis paulum incras- satis liitescentibus vel fuscescentibus ; ramis c. 5-6 mm. rarius c. 7 mm. longis, patulis vel horizontalis, strictis, obtusis, cum foliis c. 1.0- 2.0 mm. latis, simplicibiis aliquando parce (1-3) ramulosis ; caulibus ramisque dense foliosis complanatis. Folia sicca vix mutata, madida erecto-patentia vel divaricata, band homomalla ; folia dor- salia concava, obovata apice abrupte acuto-attenuata, c. 1.1-1.4 mm. longa et c. 0.6-0.72 mm. lata, asymmetrica ; folia lateralia valde concava, ovato-lanceolata, sensim in acumen angustum subu- latum strictum vel leviter curvatum attenuata, c. 1.6-1.8 mm. longa et c. 0.72 mm. lata, asymmetrica vel fere symmetrica ; folia ventralia anguste ovato-lanceolata, apice sensim anguste attenuata, c. 1.4mm. longa et c. 0.48 mm. lata; marginibus ubique integerri- mis ; nervis binis, tenuis, brevibus. c. 0.32-0.4-0.5 mm. longis ; cellulis foliorum lateralium chlorophyllosis, laevissimis, parietibus in- crassiusculis, utriculo primordial distincte, longe rhomboideis, c. 70- 100/^ longis et c. 14-16/-« latis (5-6 : 1), basilaribus infimis oblongis, c. 40/'- longis et c. 18/^ latis, marginalibus uniseriatis linearibus, alaribus indistinctis ; cellulis foliorum dorsalium superioribus late rliombeis, c. 30-40 /^ longis et c. 18-20 /^ latis. Inßorescentia autoica ; flores utriusque sexus geminiformes in caule ; folia perigonialia intima ovata, apice acuta, c. 0.4 mm. longa et c. 0.24 mm. lata, enervia ; antheridia c. 10, c. 0.2-0.24 mm. longa ; paraphysibus perpaucis. Ramulus j^ferichaetialis valde rubiginoso-radiculosus. Bra- CONTRIBUTIONES NOVAE AD FLORAM BRYOrHYTON JAPONICAM. 71 f.d. SI. A.X20 fv. ï'ig. 30. Vesiciilaria Shimadae. A. Folia : f.d. folia dorsalia ; f.l. foUa lateralia ; f.v. folia ventralia ( x 20) ; B. Sporangia ( x 10). cteae penchaetii intimae basi semivaginatae oblon- go-lanceolatao in acumen sensim. elongatum pilifor- me erectum vel cur va - tum integerrimum attenu- atae, c. 2.0-2.3 mm. longao et basi c. 0.48-0.56 mm. latae ; nervis binis, in- dictinctis, c. 0.48-0.5G mm. longis ; cellulis ut in foliis. Vag inula cylindrica, c. 0.8- 0.96 mm. longa et c. 0.4 mm. crassa, lutescens apice msca ; para- physibus paucis, hyalinis. Seta c. 13-20 mm. longa, paulum flexuosa, rubra, laevis, sicca leviter torta. TJieca pendula vel nutans, e coUo obconico distincto oblongo- cylindrica, symmetrica, c. 1.68-1.76 mm. longa (coUo c. 0.32 mm. longo) et c. 0.72-0.8 mm. crassa, fusca, laevis vel paulum mamillosa, sicca deoperculata sub ore baud cont- racta ; cellulis exothecii hexagonis vel oblongo-hexagonis. c. 30-56 x 30-40 i« magnis, ad orificiura in seriebus c. 2-3 minoribus c. 15/^ magnis hexagonis; stomatibus nuUis. Annulus duplex, c. 70/^ altus, basi luteus superne hyalinus, deciduus. Exostomii dentés subulato-lanceolati, basi paulum remoti, c. 0.48 mm. longi et basi c. 0.1 mm. lati, hyaline limbati, dorso basi lutei et dense transver- sim striatuli dein lutei et minute papillosi, apice hyalini, fere lea- ves vel paulum papillosi, linea media flexuosa, ventro hyalini dense (c. 45) et alte lamellosi ; endostomium corona basilaris lutea, minu- tissime papillosa, c. 0.16 mm. alta ; processus dentium fere longi- tudinis, carinati in carina anguste perforati, papillosi, luteo-fusci, apice hyalini ; ciha terna, bene evoluta, fere hyalina, papillosa. 72 ART. 4. — SHUTAI OKAMUEA : nodulosa. Sporae c. 9-12,« magnae, fuscescentes, laeves vel leviter scabridae. Oper cur um e basi conica acute apiculatum, c. 0.64 mm. longum et c. 0.56-0.72 mm. diam. Calyptra cuciiUata, c. 2.-2.3 mm. longa, pallida apice fusca, laevis. Formosa : Prov. Tailioku-chö, Honkeirai [^ g| ^], Shabösan [;« longis et c. 5-7 /^ latis, superioribus brevioribus c. 50-70 /^ longis, basilaribus infimis abbreviatis c. 30-40 /^ longis aureis, alaribus aureis 3-4 magnis rectangulis subvesiculosis c, 50-80/^ longis et c. 20-25 /i latis, supraalaribus quadratis c. 25/^ CONTKIBÜTIONES NOVAE AD FLORAM BRYOPHYTON JAPONIC AM. 73 raagnis, omnibus laevissimis. Inßorescentia dioica ; flores masculi ignoti. Ramuhis y^mcAaei/aZîs valde fusco-radiculosus. Bracteac pe- richaetii intimae oblongo-lanceo- latae, apice sensim anguste acuminatae, c. 24 mm. longae et c. 0.4 mm. latae, marginibus e medio ad apicem grosse serratis ; celliüis basilaribus infimis aureis. Vaglnula cylindrica, c. 1.2 mm. alta, aurea apice nigrescens ; paraphysibus numerosis, hya- linis. Seta c. 1-2 cm. longa, rubra, tenuis, laevissima, sicca torta. Theca inclinata, asym- metrica, leviter curvata, oblonga, sicca deoperculata sub ore pau- lum contracta, c. 1.8-2.0 mm. longa et c. 0.6-0.7 mm. crassa, pallido-fusca, collo brevi ; ceUulis exothecii quadratis vel rectangulis c. 28-56 x 20-28 ,« magnis, ad orificium in seriebus 1-2 transverse rectangulis c. 14x20/^ magnis; stomatibus in collo paucis. Exostomii dentés lineari-lanceolati, c. 0.4 mm. longi et basi c. 0.08 mm. lati, hyaline limbati, dorso lutei et transversim striatuli, apice lutescenti dense et grosse papillosi, ventro dense (c. 35) et alte lamellosi ; endostomium luteum ; corona basilaris c. 0.16 mm. alta, minutissime papulosa ; processus dentium longitudinis, carinati in carina anguste perforati, papillosi ; cilia unica, bene evoluta, nodulosa, hyalina, papulosa. Sporae c. 14/^ magnae, virides, minutissime et parce papillosae. Caetera ignota. Hondo : Prov. Yaraashiro ; Kamigamo, Otagi-guu (Leg. D. Imoto ! n. 1914.). PifJ. 3 1 . Rhaphidostegimn argutum. A. Planta ( x 1) ; B. Foüa ( x 25) ; C. Apex folii (xlOO); D. Sporangia (xlO). 74 AET. 4. — SHUTAI OKAMUKA : Species R. japonico Bkoth. subaffinis, sed marginibiis foliorum e medio ad apicem grosse argutis, inflorescentia dioica, seta longiore, endostomium cilia unica faciliter dignoscenda. Nome speciei ab foliis apice argutis. Rhaphidostegiinn demissmu (Wils.) De Not. Formosa : Prov. Taihoku-chô, Sliabosan (Leg. Yaichi Shimada ! 31. Xn. 1913.). Species nova ad floram formosanam. Rhaphidostegimn japonicinn Broth. Hedwigia Band XXXYin. p. 230 (1899). Habitat in truncis vel rupibus. Plant ae graciles, caespitosae, caespitibus lutescenti-viridibus, sericeis, rigidinsculis, densis, late extentis. Caulis repens, c. 3-4 cm. longiis, hie illic dense fasci- culatim fusco-radiculosus, densiiiscule subirregulariter pinnatim ramosns, sectione teres vel oblongus, c. 0.12-0.16 mm. magnus, fascicule centrali nullo, i^eti centrali luteo, cellulis tenellis hexagonis vel oblongo-hexagonis c. 10-20 /^- magnis, ceUulis periphericis 1-2-seriatis minoribus c. 10/^ magnis, valde incrassatis, fuscis ; ramis c. 6-10 mm. longis, patulis, apice obtusis, simplicibus vel parce ramulosis, dense foliosis paulum complanatis. Folia sicca adpressa rarius ad apicem rami indistincte homomalla, madida erecto-patentia, concava, e basi oblonga lanceolata, apice sensim piliformiter attenuata, c. 1.3-1.6 mm. longa et c. 0.4-0.45 mm. lata, marginibus e basi ultra medium leviter revolutis, integerrimis ; nervis nullis vel indistincte binis ; cellulis omnibus laevissimis, anguste linearibus, c. 70-80/^ longis et c. 4/^ latis, superioribus brevibus c. 35-50// longis et c. 4-6 /^ latis, basilaribus infimis brevioribus c. 40 f^ longis aureis, alaribus 3-5 magnis oblongis subvesiculosis c. 35-56 /^ longis et c. CONTRIBUTIONES NOVAE AD FLOKAM BRYOPHYTON JAPONICAM. 75 A. X20 XlO Fig. 32. Rhaphidostegium japonictun. 20-25 II Icitis aiireis, supraalaribus quad- ratis vel breviter rectangulis minoribus lutescentibiis vel hyalinis. Inflorescentia autoica ; flores utriusque sexus gemini- formes in caule vel ramis ; folia peri- gonialia intima ovata, apice breviter acnminata, c. 0.4 mm. longa et c. 0.24 mm. lata, concava, enervia ; antheridia c. 5-7, c. 0.16 mm. longa, paraphysibus a. Foiia { x 20) ; b. sporangia (xio). perpaucis, hyalinis. Ramulus perichaetialis valde fnsco-radiculosus. Bracteae perichaetii intimae oblongo-lanceolatae, apice sensim anguste attenuatae, c. 0.9-1.0 mm. longae et c. 0.24 mm. latae, subintegerrimae ; nervis binis, indistinctis ; cellulis ut in foliis. Vaginula clavellato-cylindrica, c. 0.8 mm. longa, fusca ; paraphysibns perpaucis. Seta c. 8-9 mm. longa, paulnm flexnosa, rubra, laevis, sicca torta. llicca inclinata vel suberecta, ovalis, c. 1 mm. longa et c. 0.64 mm. crassa, brevicollis, fusca, sicca deoperculata sub ore paulum contracta ; cellulis exothecii coUenchymaticis, quadratis vel breviter rectangulis, c. 30-40 ,« longis et c. 20-30 /^ latis, ad orificium in seriebus dualius vel tribus minutis hexagonis c. 15-20 /< magnis ; stomatibus nullis. Ammlus nullus. Exosfomii dentés basi connati, subulato-lanceolati, c. 0.32 mm. longi et c. 70 /^ lati, hyaline limbati, dorso e basi ad medium fusci et transversim striatuli, apice hyalini et papillosi, ventre dense (c. 35) et alte lamellosi ; endostomium luteum, minutissime papillosum ; corona basilaris c. 0.1 mm. alta ; processus dentium longitudinis, carinati in carina anguste perforati ; cilia bina vel unica, bene evoluta, nodulosa, hyalina, papillosa. Sporne c. 15/^- magnae, fuscescentes, minutissime papillosae. Operculum e basi convexo-conica rostratum, c. 0.64-0.7 mm. longum et c. 0.4 mm. diam., rostro c. 0.4-0.5 mm. 76 ART. 4. — SHÜTM OKAMUEA : longo, stricto vel curvato. Calyptra c. 1.4 mm. longa, lutescens apice fuscescens. Formosa : Prov. Taihoku-cliö ; Eifuknsan [77c ü^ Uj] (Leg. HiSAHiKO Sasaoka ! 24. n. 1913.) ; Shabösau [fflj m Oj] (Leg. Yaichi Shimada ! 30. xn. 1914.). Species nova ad floram formosanara. Distr. endemica japonica. BRACHYTHEOIACEAE. Okamtiraea brevipes Broth. Habitat in arborum truncis. Plantae graciles, caespitosae, caespitibus densiusculis, rigidiuscnlis, Intescenti-viridibua, nitidius- cnlis. Caults elongatns, repens, c. 5-ß cm. longus, divisus, subir- regulariter pinnatim ramosus, hie illic fasciculatim fusco-radicnlosns, sectione teres, c. 0.24-0.32 mm. diam., fascicnlo centrali c. 30 /^ magno paucicellulari, reti intermedio hyalino, ceUnlis hexagonis vel oblongo-hexagonis tenellis c. 15// magnis, celMis periphericis c. 5-6-seriatis sensim minoribns luteo-fuscis vel fuscis incrassatis ; ramis erectis, strictis vel substrictis, plernmque brevibus, c. 5-8 mm. longis simplicibus, rarius 13 mm. longis subpinnatim vel sub- fascicnlatim ramnlosis, omnino dense foliosis teretibus, apice obtusis vel subacutis. Folia canlina et ramea sicca imbricato- adpressa leviter plicata, madida erecto-patentia, concava, breviter decurrentia, ovato-oblonga, apice acuta, in summo acnmen sensim elongatnm subulatum angustum attenuata, c. 1.7-2.0 mm. (rarius c. 2.4 mm.) longa et c. 0.64-0.85 mm. lata, basi tri-plicata, marginibus basi late recurvis, ad apicem subplanis, integerrimis vel minutissime serrulatis ; nervo valido, ultra medium folii evanido, in sectione transversal! piano- convexo, basi c. 50 p- lato et c. 35 /^ CONTRIBUTIONES NOVAE AD FLOIîAM BRYOPHYTOX JATONICAM. 77 S.C.r. B. X25 W^' Pig. 33. Okamuraea brevipes. A. Planta (xl); B. Folia: f.c.r. folia caiüina et ramea f.rl. folia ramiilina {x25); C. Sporangia (xlO). XlO crasso, e 4-stratis cel- lularum uniformium composito ; cellulis lae- vissimis, plus minus incrassatis, rhombeis vel rhomboideis, c. 20-35 It longis et c. 6-9 /^ latis, superiori- bus longioribus angus- tioribus, basilaribus quadratis breviter rec- tangulis breviter linearibus vel oblongo-hexagonis c. 20-25 /^ longis et c. 9-14 [J- latis, alaribus numerosis quadratis breviter rectangulis c. 9-14 /i longis et c. 7/^ latis; folia ramulina angustiora, anguste oblongo-lanceolata, apice sensini attenuata. Inflorescentia autoica ; flores utriusque sexus geminiformes in caule vel ramis ; folia perigonialia intima ovata apice acuto-attenuata, c. 0.64 mm. longa et c. 0.32 mm. lata, nervo tenui, ad medium evanido ; antheridia c. 5-6, c. 0.4 mm. longa, paraphysibus numerosis, hyalinis. Ramulus per ichaet talis hand radiculosus. Bracteac perichaetii intimae e basi vaginante sensim anguste subulatae, subula erecta inte- gerrima, c. 1.9 mm. longae et c. 0.56 mm. latae ; nervo tenui, medium evanido ; cellulis linearibus. Vaginida cylindrica, c. 0.8 mm. longa, lutescens apice fusca ; paraphysibus sat numerosis lutescentibus vel hyalinis. Seta c. 3-5 mm. longa, stricta curvatula vel leviter flexuosula, rubra, laevissima. Theca erecta, symmetrica, oblonga, c. 1.2-1.4 mm. longa et c. 0.7 mm. crassa, brevicolhs (collo obconico, c. 0.3 mm. longo), castanea, laevissima ; celluHs exothecii quadratis rectangulis vel hexagonis, c. 20-45/^ longis et c. 20-30 A« latis, ad orificium in seriebus c. 4-6 minutis hexagonis ; 78 ART. 4. — SHUTAI OKAMURA : stomatibus in collo numerosis, phaneroporis. Annulas duplex, c. 20/^ altus, fuscus, deciduus. Exostomii dentés basi connati, lineari- lanceolati, c. 0.28 mm. alti et basi c. 50 /^- lati, linea media fle- xuosula, marginibus dense cristatis, strato dorsali Inteo minntissime papilloso, strato ventrali lato albido vel lutescenti densissime (c. 40) et alte lamelloso ; endostomium hyalinum, minntissime papillosum vel laeve ; corona basilaris humilis, c. 50 /^ alta ; processus nnlli ; cilia nulla. Sporae c. 15/^ magnae, lutescenti-virides, subscabridae. Operculum e basi conica longe rostratum, c. 0.9-1.0 mm. longum et c. 0.4 mm. diam., rostro longiusculo, recto, apice obtuso. Calyptva ignota. Hondo : Prov. Ecliigo ; in monte Ogino jö-jama [^ / ttS< |ij] (Leg. Masaö Nakamura! 15. V. 1908.). Okamuraea cristata Broth, in Öfvers. Finsca Vet.- Soc. Förh. XLIX. No. 10. p. 2. (1905/06) ; Bot. Magaz. Tokyo. Vol. XXII. No. 254. pp. 41-43 (1908) ; Engler und Prant. nat. Pflanzf. p. 1132. (1908). Shikoku: Prov. To- sa ; in monte Tsuetate-töge (Leg. Shütai Okamura ! 6. IV. 1906. et 29. Xn. 1906); in monte Yokogura (Leg. Shütai Okamura ! 26. V. 1907.) ; in monte Iradu (Leg. Shütai Okamura ! 1. I. 1908.). Hondo: Prov. Tan- go ; Nariaisan (Leg. ICHIRO Sono ! 28. XI. 1907.). Prov. Rikazen ; Kinkwazan (Leg. ÉJIRO Uematsu ! 31. in. 1908.). Prov. Shinano ; in monte Tateshina (Leg. Zenryô Öhinata ! 12. VI- 1910.) : in monte Asama (Leg. Zenryô Öhinata ! 20. IV. 1912.). /er. A. x25 frl. Fig, 34. Okamuraea cristata. A. Folix : f .er. folia cauliiia et ramea ; f .rl. folia ramialina ( x 25). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 79 Okamuvaea cvistata Broth, var. gracilis Broth. Habitat in truncis arborum. Plantae graciles, caespitosae, caespitibus late extentis, densis, rigidiusculis, lutescenti-viridibus, nitidiiisculis. Caulis elongatus, repens, usque ad 4 cm. longus, divisus, subpinnatim ramosus, hie illic fasciculatim fusco-radiculosus, sectione teres, c. 0.16-0.2 mm. diam., fascicule centrali c. 28/^ magno pauci-cellulari, reti intermedio hyaline, cellulis hexagonis vel oblongo-hexagonis tenellis c. 9-14/^ magnis, cellulis periphericis c. 3-4-seriatis sensim minoribus lutescenti-fuscescentibus incrassatis ; ramis erectis, strictis vel curvatulis, plerumque brevibus, c. 5-12 mm. longis, simplicibus vel subpinnatim vel subfasciculatim ramulosis, dense foliosis teretibus, apice obtusis vel breviter attenuatis. Folia caulina et ramea sicca adpressa, madida erecto- patentia, concava, basi tri-plicata, breviter decurrentia, late ovato- lanceolata, apice sensim elonga- te subulate attenuata, c. 1.2 mm. longa et c. 0.56 mm. lata, marginibus basi late recur vis dein ad apicem subplanis, integerrimis vel minutissime serrulatis ; nervo valido, ultra medium folii evanido, in sectione transversah plano- convexo, basi c. 42/^ lato et c. 28/^ crasso, e 4-stratis cellularum incrassatarum uniformium composite ; cellulis laevissimis plus minusve incrassatis, in medio fohi rhombeis vel oblongis, c. 15-17/^ longis et c. 7-8 /j- latis, superioribus longioribus rhomboideis vel sublinearibus c. 20-30 /^- longis et c. 6-7 /^ latis, basilaribus quadratis oblongis vel breviter rectangulis c. 10-20// longis et c. 7-10/^ A.X25 Fig. 35. Okamuraea cristata var. gracilis. A. Folia ; f .c.r. folia caulina et ramea ; f .rl. folia ramiilina ( x 25). 80 ART. 4. — SHÜTAI OKAMUBA : latis, alaribus numerosis quadratis c. 7-10 // magnis ; folia ramulina late ovata apice subito subulato-attenuata. Hondo: Prov. Ise; Tsü-shi (Leg. Hisahiko Sasaoka ! 13. YI. 1913). A typo statura graciliore, foliis minoribus, apice brevioribus differt. Okainuraea Jiakoniensis (Mitt.) Beoth. Engler und Pbantl nat. Pflanzenf. p. 1133. (1908). Syn. Hypmifii liakoniense Mitt. Trans. Linn. Soc. Lond. 2iid Ser. Bot. Vol. ni. Part 3. p. 185. (1891.). Habitat in arborum truncis. Plcmtae graciles, caespitosae, caespitibus late extentis, densiusculis, rigidiusciilis, lutescenti- viridibus, nitidiusculis. Caulis elongatus, repens, usque ad 5 cm. longus, irregulariter pinnatini ramosus, hie illic fasciculatim fusco- radiculosus, sectione teres, c. 0.24 mm. diam., fascicule centrali pauci-cellulari, reti intermedio hyaline, cellulis hexagonis vel oblonge -hexagonis tenellis c. 14/^ magnis, cellulis periphericis c. 5-6-seriatis minoribus lutes- centibus vel fuscis valde in- crassatis ; ramis erectis, strictis vel curvatulis, c. 2 cm. longis, subpinnatim vel subfasciculatim ramulosis vel simplicibus, dense foliosis teretibus, apice attenuatis saepe flageUiformiter productis ; ramulis attenuatis. Folia cauhna et ramea sicca adpressa, madida erecto-patentia, concava, vix decurrentia, ovato-oblonga, in acumen sensim elongatum subulatum attenuata, c. l.G-1.8 cm. longa et c. y-C-r. A.X25 Fiff. 36. Okamuraea hakoiiiensis. A. Folia : f.c.r. folia caulina et ramea f.rl. folia ramulina ( x 25) ; B. Sporangia ( x 10). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 81 0.5-0.64 mm. lata, basi tri-plicata, marginibus basi late recm^vis dein ad medium siibplanis, integerrimis vol minutissime serrulatis ; nervo valido, ultra medium folii evanido, in sectione transversali plano-convexo, basi c. 42 /^ lato et c. 30 !>■ crasso, e 3-stratis cellulalum incrassatarum uniformium composite ; cellulis laevissimis, plus minusve incrassatis, rhombeis vel rliomboideis, c.20-35 /^ long- is et c. 7-9 /^ latis, superioribus longioribus angustioribus, basilaribus late rliomboideis rectangulis vel breviter linearibus, alaribus nu- merosis minoribus quadratis vel breviter rectangulis (lumine saepe ovalibus) vel oblongis c. 7-9 /^ magnis ; folia ramulina angustiora, oblongo-lanceolata, apice anguste attenuata. Inßorescenüa dioica ? floribus masculis hand visis dioica videtur. Ramulus pericJiaetialis baud radiculosus. Bracteae perichaetii intimae e basi vaginante sensim anguste subulatae, subula erecta vel leviter curvata, integer- rimae, c. 1.6-2.0 mm. longae et c. 0.4 mm. latae ; nervo tenui, medium evanido ; cellulis linearibus, basilaribus rectangulis. Vagi- nula cylindrica, c. 0.8-1.0 mm. longa, lutescenti-viridescens apice fusca ; parapliysibus sat numerosis, lutescentibus vel hyalinis. Seta c. 7-9 mm. longa, stricta curvatula vel leviter flexuosula, rubra, laevissima, sicca torta. Theca erecta vel inclinata, breviter oblonga vel ovalis, symmetrica vel paulum gibbosa, c. 1.2-1.5 mm. longa et c. 0.64-0.72 mm. crassa, castanea, laevissima, nitidiuscula, brevicoUis, coUo obconico c. 0.24-0.3 mm. longo ; cellulis exothecii quadratis hexagonis vel rectangulis, c. 30-50 ,"■ longis et c. 20-35 // latis, ad orificium in seriebus c. 4-5 minutis hexagonis c. 20// magnis ; stomatibus in coUo numerosis, phaneroporis. Annulus duplex, c. 28-30 ,« altus, ex miica série ceUularum, fuscescens, deciduus. Exostomii dentés basi connati, lineari-lanceolati, c. 0.32-0.36 mm. alti et basi c. 56 f- lati, linea media flexuosula, marginibus dense cristatis, strato dorsali luteo minutissime pa- 82 ART. 4. — SHUTAI OKAMURA : pilloso, Strato vontrali angustiore luteo densissime (c. 35-40) et alte lamelloso ; endostomium hyalinum, minntissime papillosum ; corona basilaris c. 0.1mm. alta ; processus nulli ; cilia nulla. Sporae c. 28-35 /-« magnae, virides, subscabridae. Operculum e basi conica longe rostratum, c. 1.0-1.2 mm. longum, rostro longo, subulato, curvatulo, apice obtuso. Calyptra cucuUata, c. 1.7 mm. longa, lutescens apice fusca, parce pilosa, pilis hyalinis. Shikoku : Prov. Tosa ; in monte Yanaze (Leg. Shütai Okamura ! 20. YTTT. 1905.). Hondo : Prov. Shinano ; in monte Yatsuga-dake (Leg. Ëjiro Uejviatsu ! 24. Vm. 1908.). Prov. Idu ; in monte Amagi (Leg. Kiyoshi Füjn ! 18. vm. 1914.). Prov. Suruga ; Gotenba (Leg. Kiyotaka Hisauchi ! 11. IV. 1915). Okainuraea imbricata Broth. Habitat in truncis arborum. Plantae graciles, caespitosae, caespitibus late extensis, densiusculis, rigidiusculis lutescenti-viridi- bus, nitidiusculis. Caulis elongatus, repens, usque ad 7 cm. longus, divisus, irregulariter pinnatim ramosus, hie illic fasciculatim fusco- radiculosus, sectione teres, c. 0.28 nmi. diam., fascicule centrali pauci- ceUulari, reti intermedio hyaline vel lutescenti, ceUulis hexagonis teneUis c. 15// magnis, cellulis periphericis c. 5-6-seriatis minori- bus fuscis valde incrassatis ; ramis erectis, strictis vel curvatuHs, usque ad 2 cm. longis, subpinnatim vel subfasciculatim ramulosis vel sim- plicibus, dense foliosis teretibus, apice obtusis vel attenuatis. Folia caulina et ramea sicca imbricato- f.rl. A. X25 PifJ- 37. Okamuraea imbricata. A. Folia : f. er. folia, caulina et ramea ; f.rl. folia ramiilina ( x 25). CONTRIBUTIONES NOVAE AD FLOKAM BRYOniYTON .JAP0NICA:\[. 83 adpressa, inadida erecto-patentia, concava, brovitor decurrcntia, ovato-lanceolata, in acumen sensim subulatum attenuata, c. 1. 3(3-1. (> mm. longa et c. 0.64-0.72 nnn. lata, basi tri-plicata, marginibus basi late recurvis dein ad medium subplanis, integerrimis vel minutissime serrulatis ; nervo valido, infra apicem folii evanido, in sectione transversal! piano- con vexo, basi c. 35// lato et c. 20/''- crasso, o 3-4-stratis cellularum incrassatarum uniformium composito ; ceUulis laevissimis, plus minusve incrassatis, rhombeis vel rhomboideis, c. 20-30 !>■ longis et c. 6-8 /^ latis, superioribus longioribus angustiori- bus, basilaribus rhomboideis rectangulis vel breviter linearibus, alari- bus numerosis minoribus quadratis (lumine saepe ovalibus) c. In magnis ; folia ramulina angustiora, ovato-oblonga, apico attenuata. Caetera ignota. Hondo : Prov. Mutsu ; in monte Hakkoda (Leg. Ëjibô Uematsu ! 31. VII. 1907.). Prov. Rikuzen ; Ourabayaslii, Sendai (Leg. Ejirö Uematsu ! 27. IX. 1907.). Okaniuraea tissuriensis (Beoth.) Broth. Engler und Peant. nat. Piianzenf. p. 1133. (1908). Syn. Bryhnia ussioiemis Broth. Frag. Hor. bryolog. Asiae orient. cong. IL pp. 8-9. (1906). Habitat in saxis. Plantae robustiusculae, caespitosae, caespiti- bus lutescenti-viridibus, nitidiusculis, rigidiusculis, densis, late extentis. Caulis elongatus, repens, usque ad 4 cm. longus, irregu- lariter pinnatim ramosus, rarius fusco-radiculosus, sectione teres, c. 0.24 mm. diam., fascicule centrali valde arto pauci-ceUulari, reti intermedio hyalino, cellulis hexagonis tenellis c. 10-15/^ magnis, cellulis periphericis c. 3-4-seriatis minoribus lutescenti-fuscescentibus incrassatis ; ramis prostratis vel ascendentibus, usque ad 2.5 cm. longis, irregulariter ramulosis ; ramis ramulisque dense foliosis, api- 84 AET. 4. SHCTAI OKAMURA ce obtusis acutis vel attenuatis. ParaphyUia nulla. Folia caulina et ramea sicca adpressa, madida erecto-patentia, concava, breviter decurrentia, late cordato-ovata, apice acuta inde subito in summo acumen elongatum angustum subulatum attenuata, c. 1.9 mm. longa et c. 0.9 mm. lata, basi tri-plicata, marginibus late recurvis, integerrimis vel minutissime serrulatis ; nervo vaUdo, infra apicem folii evanido, in sectione transversali piano -convexo, basi c. 56 /< lato et c. 35 /^- crasso, e 4-stratis ceUularum incrassatarum uni- formium composite ; ceUulis laevissimis, plus minusve incrassatis. B.X15 Fig. 38. Okaniuraea ussiiriensis. A. Planta ( x 1) ; B. Folia : f.c.r. folia caulina et raniea ; f.rl. folia ramulina ( x 15). rhombeis vel rhomboideis, c. 15-20/^ longis et c. 5-8/^ latis, superioribus longioribus, basilaribus quadratis c. 7-10/^ magnis vel rectangulis vel rhombeis c. 20 /^ longis et c. 4 /^ latis, alaribns numerosis minoribus quadratis rectangulis vel rhombeis ; folia ramuhna angustiora, oblongo-lanceolata, apice sensim attenuata. Caetera ignota. Korea : Prov. Heian-HokudO ; in monte Hiraihö [^ jR P)^] (Leg. TxVKe- NosmN Nakai ! 9. YI. 1914.). Species nova ad floram koreanam. Distr. Ussuri et China. 1. 4. /• CONTRIBUTIONES NOVAE AD FLORAM BEYOPHYTON JAPONICAM. 85 Conspectus specierum et varietatum Okamuraearum. Folia caulina et rainea apice subito elongate attennata 1. Folia ca,ulina et raraea apice sensini breviter vel longe attenuata. 2. Folia caulina et ramea ovato-oblouga, apice snbrotundato-acuta inde subito iu sumrao acumen elougatum subpiliforme attenuata. 0. plicafa. Folia caulina et ramea late cordato-o^ata, apice acuta inde subito in summo acumen elongatura angustum subulatum attenuata 0. ussuriensis. Folia caulina et ramea late ovato-lanceolata, apice breviter subulato- 2. { attenuata 0. imhricata. Folia caulina et ramea apice longe attenuata 3. Folia caulina et raraea late ovato-oblonga, apice subutato-attenuata ; seta c. 3-5 mm. longa 0. brevipes. Folia caulina et ramea ovato-lanceolata, apice elongate et subulate attenuata ; seta c. 7-15 mm. longa 4. Folia caulina et raraea anguste ovato-lanceolata ; seta c. 7 -9 ram. longa ; theca breviter oblonga vel ovalis, c. 1.2-1.5 mm. longa. 0. hakoniensis. Folia caulina et ramea late ovato-lanceolata ; seta c. 15 mm. longa ; tlieca oblonga c. 2 mm. longa 0. ciistata. i Ramis raraulisque longissirae raulti-flagelliferis ; caetera ut in typo. I 0. crisf. var. mnUtflagélliferoa. Rarais plerumque brevibus hand flagelliferis ; folis minoribus, apice brevioribus ; planta graciliore 0. crist. var. gracilis. Ishibaea jajwnica Broth, et Sh. Okamura. Bot. Magaz. Tokyo. Vol. XXIX. No. 346. pp. 186-188. Tab. YIH. (1915). Habitat in truncis arborum. Plantae tenellae, caespitosae, caespitibus latescenti-viridibus, nitidiusculis, rigidinsculis. Caulis repens, usque ad G cm. longus, hie illie faseieulatim rul^iginoso- 86 ART. 4. — SHUTAT OKAMURA : radiculosns, densiusculo et iiTogularitor raiiiosus, sectiono toros, e. 0.2-0.24 iTiiu. diam., fasciculo contrali niillo, roti centrali hyalino, cellulis iiTogularitor hoxagoniB (parietibus flexiiosis) c. 20/^- iiiagnis, cellulis poriplicricis Intescentibus vcl fuscis c. 2-3-seriatis minoribus incrassatis ; ramis prostratis vel paulum asccndontibus, simplicibus vel parc(i (l-o) rainiilosis, strictis vel leviter curvatiö, apice obtiisis vel acutis, usque ad 1 cm. longis, dense foliosis et teretibus. ParaphyUia nulla. Folia sicca adperessa, madida ereeto-patentia, valde concava, ovato-lanceolata, in acumen scaisini subulatuni attenuata, c. 1.2-].-) ]nni. longa et c. 0.28-0.32 mm. lata, margini- bus usqu(^ ad basin acuminis rc^volutis, apice serrulatis vel sub- integerrimis ; nervo infra summum apicem folii evanido, in sectione transversali plano-convexo, dorso valde prominent!, basi c. 42 /^ lato et c. 30 /'- crasso, e 4-stratis cellularum uniformium incras- satar^ni comi)Osito, cellulis ventralibus 4-5 majoribus, dorsalibus c. 9 niinoribus ; cc41ulis laminalibns prosenchymaticis, laevibus, valde chloropliyllosis, c. 30-r)(;// longis et c. 3-5/^ latis, superioribus brevioribus, basilaribus laxioribus, angularibus sat numerosis quad- ratis c. 9 1>- magnis, chloropliyllosis. Inßorescenfia autoica ; flores masculi in caule geminiformes ; folia perigonialia externa obovato- oblonga, apic(^ longe attenuata, intima obovato-oblonga, apice acuta, c. ().(>4 nmi. longa et c. 0.24 mm. lata, concava, enervia ; antheridia c. 0, ])araphysibus paucis, c. 0.24 nim. longis, liyalinis ; flores feminei in ramis rarius in caule positi. Ramulus perichaetialis parce radiculosus. Bradeae perichaetii intimae basi semivaginatae oblongo-lanceolatae, subito in acumen elongatum subulatum attenuatae, c. 1.6 mm. longae, apice minutissime serrulatae ; nervo infra sumnaum apicem evanido ; cellulis linearibus, basilaribus longe rectangulis. Vaginula cylindrica, c. 0.8-0.9 mm. longa, albicanti- fuscescens, c. 1.2-1.5 mm. alta. Srfa stricta vel leviter flexnosula, CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAIM. 87 rubra, lac'vis, sicca snporno torta. Theca orocta, oblonga vol oblongo-cylindrica, syininotrJca, c. 1.2-1. G mm. Tonga et c. 0.56-0.72 mm. crassa, lutoo-rnbra, laevis, collo conico brevi ; ccUulis exothocii irrogularitcT rectangiilis vol qnadratis, ad orificiuTii in scricbus c. 6 minoribuK liexagonis ; stomatibus in collo paucis phancroporis. Annidns nullis. Peristom'mm dui)lox ; exostomii dontcvs subulato- lanceolati, basi connati, c. 0.2-0.24 mm. longi ct basi c. 42-50/^ lati, strato dorsali luteo vol lutescenti-fusco, basi transversim striolato dein ad apicem laevissimo, linea media llexuosula, strato ventrali luteo angustioro c. 18-20 lamcUoso ; endostorainm flaves- cens, laeve ; corona basilaris c. 40 /^ alta ; processus dentium 4^5 longitudinis, leviter carinati, in carina perforati, sed plerumque divisi et fragosi, fragmentis dentibus adhaerentibus ; cilia nulla. Sporae c. 15// magnae, fuscescenti-virides, laeves. Operculum conicum, apice obtusum, c. 0.32 mm. longum et c. 0.32 mm. diam. Cahjptra cucuUata, c. 1.9 mm. longa, lutescenti- vi vi descens apico fusca, laevis. Hondo : Prov. Shinano ; Sironmajiri (Leg. Ëkichi Ïshiba ! 17. VIII. 1909.) ; Sakai-mura, Shimotakai-gnn (Leg. Shinzô Ïtô ! 28. X. 1913.). Genus Homalothecio Be, euk. affine, sed inflorescentia, tenoritato omnium partium, foliis hand plicatis nee non peristomii structura exquo longe diversum. Genus clar. Ekichi Ishiba, florae bryologicae Japoniae scruta- tor! meritissimo dedicatum. Ptychodiiim pevatt&tiiiatu/ni Sii. Okamuea. sp. no^ . Habitat in truncis arborum. Plcmtae robustiusculats denser caespitosae, late extentae, lutescenti-virides, nitidae, rigidiusculae. Calais repens, elongatus, usque ad 10 cm. longus, hie illic dense fasciculatim fusco-radiculosus, dense et regulariter vel subregnlariter 88 ART. 4. — SHÜTAI OKAMUEA pinnatim ramosus, sectione oblongns, c. 0.28-0.32 x 0.2-0.24 mm. magnus, fasciculo central! nnllo, reti centrali liyalino, cellulis hexagonis tenellis c. 15/>- magnis, cellulis periphericis c. 4-5-seriatis minoribus incrassatis lutescentibus vel fuscescentibus ; ramis ads- cendentibus, c. 5-15 mm. longis, strictis vel curvatis, simplicibus vel parce (1-4) fasciculatim rarins pinnatim ramulosis, apice obtusis acntis vel attenuatis, aliquando flagelliformibus, dense foliosis, teretibus. Paraphyllia nuUa. Folia sicca laxe adpressa plicata hand liomomaUa, madida erecto-patentia, concava, sub- carinata, longitudinaliter plicata (c. 5), basi breviter decurrentia, late ovato- vel subcordato-deltoidea subito in acumen elongatum angustum subulatum summo apice subloriforme attenuata, c. 2.7-2.9 mm. longa et c. 1.0-1.2 mm. lata, marginibus basi revolutis dein subplanis, e basi ad apicem minute saepe in acumen grosse serratis ; nervo validiusculo, infra summum apicem folii evanido vel fere percu- rrente, in sectione transversali piano - convexo, dorso prominenti, basi c. 50,« lato et c. 20/>< crasso, e 4-s- tratis cellularum incrassatarum uni- formium composito ; cellulis laevibus, parietibus incrassiusculis, linearibus, c. 30-40/^ longis et c. 5-6/^ latis, basilaribus c. 40-70 /^ longis et c. 6-7 /^ latis, infimis brevioribus et latioribus c. 9/^ latis parce porosis, alaribus numerosis breviter rectangulis c. 20-35 x 9-15 /^ magnis hyalinis, supraalaribus numerosis quadratis c. 15/^ magnis vel oblongis c. 20,« longis hyaünis. In- ßorescentia dioica ; flores utriusque sexus geminiformes in caule et Fig. 39. Ptychodium perattemiatiini. A. Folium ( X 20) ; B. Theea ( x 10). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JArONICAM. 89 ramis ; folia perigonialia intima ovata, apice subito subulate attenuata, c. 0.8 mm. longa et c. 0.3.") mm. lata, nervo tenui, influa apicem eva- nido; antlieridia c. 12, c. 0.4 mm. longa, paraphysibus numerosis, liyalinis, c. 0.4 mm. longis. Eamulus pcrichaetialis parce radicnlosus. Bracteae pericliaetii intimae e basi alte semivaginante subito in acumen elongatum angustum loriforme attenuatae, apice recurvae serratae, c. 2.8-3.2 mm. longae et c. 0.64-0.72 mm. latae ; nervo tenui, infra summum apicem evanido ; cellulis ut in foliis, sed basilaribus laxioribus fere anguste rectangulis. Vaginula cylindrica, c. 1.4-1.6 mm. alta et c. 0.3 mm. crassa, pallida ; paraphysibus numerosis, liyaKnis vel lutescentibus. Seta c. 13-15 mm. longa, stricta vel leviter flexuosa, rubra, laevissima, sicca torta. Theca erecta, oblonga vel oblonge -cylindrica, symmetrica, c. 2.0-2.4 mm. longa (coUo c. 0.32-0.4 mm. longo) et c. 0.8 -O.Î)6 mm. crassa, fusca, laevis ; cellulis exothecii hexagonis oblonge -hexagonis vel rectangulis, c. 20-50x20-25," magnis, ad orificium in seriebus c. 2-3 minori- bus c. 1 5 fj- magnis quadratis vel hexagonis ; stomatibus in collo sat numerosis. Annidus duplex, c. 40 ,« altus, fuscescens. Exostomii dentés hneari-subulati, basi connati, c. 0.4 mm. longi et basi c. 50-56,« lati, dorso lutei, basi dense transversim striatuli, superne hyalini minute papillosi, linea media flexuosa, ventre dense (c. 30) et altiuscule lamellosi ; endostomium corona basilaris c. 0.1 mm. alta, lutea, minute papiUosa ; processus dentium fere longitudinis, carinati in carina hand vel angustissime perforati, lutei, minute papillosi ; cilia unica, brevia vel rudimentaria. Sporae c. 9 — 11,« magnae, luteae, laeves. Operculum e basi anguste conica longe rostratum, c. 1.2-1.6 mm. longum et basi c. 0.56 — 0.7 mm. diam., rostre recto vel leviter curvato, apice obtuse. Calyptra cucullata, c. 2.4-2.8 mm. longa, pallida apice lutea, laevis. 90 AET. 4. SHÜTAI OKAMUEA : Formosa : Prov. Taihoku-chô ; Shabôsan [fi^ ift Ui] (Leg. Yaichi Shi- MADA ! 30. XII. 1914.). Species P. jjlicahùlo Caed. valde affinis, sed foliis longioribiis, apice subito elongate et anguste atténua tis di versa. Nomen speciei ab foliis perattenuatis. JBrachytheciuin plumosum (Sw.) Bii. eue. Formosa : Prov. Taihokn-chô ; Sliabôsan (Leg. Yaichi Shimada ! 2. I. 1913.). 3Iyuroclada concinna (Wils.) Besch. Korea : Prov. Kankyô-nando ; Shôshinzan [fô E Uj] (Leg. Takenoshin Nakm ! 27. Vm. 1914.). OxyrrhynchiuTti Schotttn'ùlleH (Broth.) Broth. Hondo : Prov. Ivôduke ; Maru-numa (Leg. Haraikaava ! "VII. 1914). Habitat in solio lacus c. 2 m., prof undo. Oxyrrhyuchiuin Schottniiïllet'l (Broth.) Broth, var. perlongicladum Sh. Okamura. var. nov. Habitat in rivulornm saxis irroratis. Plantac robustae, caes- pitosae, caespitibus late extentis, lutescenti-viridibus, nitidiusculis, rigidiusculis. Caidis repens, valde elongatus, c. 25 cm. longus, basi nudus, baud radiculosus, densiuscule ramosus, sectione sub- angulato-teres, c. 0.5 mm. diam., fasciculo centrali c. 40 /^ magno pauciceUulari, reti intermedio hyaline, ceUulis hexagonis c. 25- 30/^ magnis, cellulis periphericis c. 5-seriatis minoribus lutescenti- fuscis subincrassatis ; ramis repentibus, perelongatis, usque ad 18 cm. longis, subflexuosis, laxiuscule foliosis haud complanatis et hic illic nudis, apice obtusis vel breviter attenuatis, vage ramulosis ; CONTKIBUTIONES NOVAE AD FT.OIÎA3I BRYOPHYTOX J.VPONICAlSr. 91 ramnlis patiilis, 1-3 cm. longis, laxiuscnlo foliosis, hand complanatis. Folia sicca patula vol laxc adprossa ct paiilum nintata, madida erocto-patentia, concaviusciila, brovitor decurrentia, o basi angusta late ovata vel ovato-ovalia, apico obtusa vol obtiisinscula, c. 1.4-1.0 mm. longa ot c. 1.3-1.0 mm. lata, marginibus basi revolutis doin planis, fore ubique sorrnlatis ; norvo longo nltra modinm ovanido ; collnlis linearibus, laevissimis, valdo chlorophyllosis, c. 70-112// longis ot c. 8-0 /^ latis, snporioribns brevioribns rbomboidois c. 20-28 li longis t^t c. 8-10 /^ latis, basilaribus infimis brovioribns laxioribus c. 12-14// latis, alaribus roctangulis c. 30-00// longis et c. 15-20// latis fuscis vol chlorophyllosis. Bracteac pcrichnetii e basi doltoideo-ovata sensim acuminatae, acnmine canalicnlatao, c. 1.4-1.0 mm. longae ot c. 0.50-0.04 mm. latao, concavae, marginibus e medio ad apicom serratis ; nervo tonui, infra apicem evanido. Seta c. 10-12 mm. longa, flexuosula, tennis, rubra, laevissima, leviter torta. Theca inchnata, oblongo-cylindrica, symmetrica, c. 2 mm. longa et c. 0.8 mm. crassa, fusca, laevis. Sporac virides, c. 1 0-17 ff- magnao, laeves. Caetera ignota. Hondo : Prov. Musashi, Ivobotoke-toge (Leg. Kiyotaka Hisauchi ! 20. IN. 1914.). A tipo ramis porelongatis, hand complanatis differt. Nomon varietatis ab ramis porelongatis. EurrJiyncJiinm Arbiismila Broth. Kyushu : Prov. Ösumi, in insula Yakushima (Leg. Takanori Iwaki ! 21. IX. 1914.). Etirrhynchium yexoanu'ni 8h. Okajiura sp. nov. Habitat in terra. Plantae gi*aciles, caespitosae, caespitibus viridibus vel fusco- vol lutoo-vii-idibus, nitidiuscuhs, rigidiusculis, late 92 ART, 4. — SHÜTAI OKAMURA extentis, densis. Caulis repens, elongatiis, usque ad 4 cm. longus, dense fasciculatim fusco-radiculosus, densissime et regulariter pinnatim ramosus, sectione teres vel oblongus, c. 0.3-0.4 mm. diam,, fasciculo centrali arto paucicellulari, reti intermedio hyalino, cellulis hexagonis c. 14-20/^- magnis, cellulis periphercis c. 4-5-seriatis minoribus c. 7-10/^ magnis paulum incrassatis lutescentibus ; ramis erectis suberectis vel ascendentibus, c. 5-10 mm. longis, inferne simplicibus stipitiformibus dense fasciculatim fnsco-radiculosis, super- ne dense pinnatim ramulosis, apice obtusis ; ramulis erectis, c. 2-4 mm. longis, apice obtusis ; ramis ramulisque dense foliosis teretibus. Paraphyllia nulla. Folia sicca laxe adpressa, madida erecto-patentia, dimorpha ; folia cauUna breviter decurrentia, e basi cordata vel deltoideo-cordata breviter lanceolato-attenuata, apice saepe recurva, c. 0.48-0.65-0.8 mm. longa et c. 0.4-0.45-0.48 mm. lata, marginibus basi late recnrvis, ubique serrulatis ; nervo tenui, infra apicem folii evanido, superne sensim tenuiore ; cellulis valde chlorophyUo- sis laevibus, linearibus, c. 28-50 >« longis et c. 3-4 /^ latis, superioribus angustioribus, basilaribus brevioribus ot laxioribus c. 14-20/^- longis et c. 7 />« latis, alaribus numerosis quad- ratis vel rectangulis c. 10-14/^ longis et c. 10/^- latis; folia ramea valde concava, vix decurrentia, cordata vel ovato-cordata, apice breviter acuta obtusa vel rotundato-obtusa, c. 0.32-0.4 mm. longa et c. 0.28-0.32 mm. lata, marginibus basi late recurvis, ubiqne serratis ; nervo valido, infra summum apicem folii evanido, apice Fig. 40, Eurrhynchium yezoaiium. A. Folia: f.c. folia caulina, f.r. folia xamea (x40); B. Sporangia (xlO). CONTEIBUTIONES NOVAE AD FLORAM BRYOPIIYTON JAPONICAM. 93 hand tenuiore et extremitato dorso denticulo instriicto, in sectione transversal! piano -convexo, dorso valde prominenti, basi c. 45 ff- crasso, e 5-stratis cellularum incrassatariim uniformium composito ; cellulis chlorophyllosis, laevibus, linearibus, c. 25-35 /'- longis et c. 3-4/^ latis, superioribus brevioribus c. 18-20/^- longis basilaribus brevioribns et laxioribns c. 14-20 /^ longis et c. 5-7 /^ latis, alaribns qnadratis vel rectangulis c. 8-14/^- longis et c. 6-8,« latis con- coloribus. Inflorescentia dioica ? (floribns mascnlis hand visis dioica videtur). Hamulus perichaetialls radicnlosus. Bracteae perichaetii intiraae basi semivaginatae ellipticae, subito in acnmen sensim elongatnni angnstum lineari-subulatum attenuatae, c. 1.4-1.8 mm. longae et c. 0.4-0.5 mm. latae, acumine fere erectae, remote serrulatae ; nervo nnllo ; cellulis linearibus c. 80-90 ,« longis et c. 4-5 ,« latis, basilaribus brevioribus et laxioribns c. 40-70 /^ longis et c. 9.-14/'- latis. F«^mzi/a cylindrica, c. 10 mm. alta, albescenti- lutescens apice fusca ; archegonia numerosa ; paraphysibus numéro- sis, hyalinis. Seta c. 6-8 mm. alta, erecta vel leviter curvata, rubra, laevissima, sicca hand torta. Theca inclinata, e collo brevi oblonge -cylindrica, asymmetrica, paulum gibbosa, sicca deoperculata sub ore contracta, c. 1.2-1.6 mm. longa et c. 0.6-0.7 mm. crassa, lutescenti- rubra ; cellulis exothecii qnadratis hexagonis vel rectangulis c. 28-40x14-20/^ niagnis, ad orificium minoribus rotundato- hexagonis c. 10-14 r- magnis ; stomatibus in collo sat numerosis, phaneroporis. Aimulus c. 30-35 /^- altus, simplex, e 1 vel 2 seriebus cellularum, lutescenti-ruber, deciduus. Pristomium duplex ; ex- ostomii dentés subulato-lanceolati, basi connati, c. 0.4-0.48 mm. longi et basi c. 70/^ lati, strato dorsali basi luteo vel fuscescenti- luteo et transversim striatulo, e medio ad apicem lutescenti et papillose, strato ventrali angustiore dense (c. 30) et alte lamelloso ; endostomium luteum ; corona basilaris c. 0.18 mm. alta, fere laevis ; 94 ART. 4. — SHUTAI OKAMÜRA : processus deiitium fere longitudinis, carinati in carina angustissime perforati vel haud perforati ; cilia unica, brevia, fere laevia, nodu- losa, latioria, aliquando bina et angustiora. Sporae c. 12-17 mm. magnae, luteo-virides laeves. Opereidum e basi conica longe et paulum curvate rostratum, c. 0.8-1.0 mm. longum et c. 0.48 mm. diam., lutescenti-rubrum, rostro apice obtuso laevi. Ccdyptra cucuUata, Kiteo-viridescens, nuda, laevis. Yezo (Hokkaido) : Prov. Isliikali, Yamaliana-mura, Sapporo (Leg. Nizö IwASAKi! 21. IV. 1914.). Species cum E. strigoso (Hoffm.) Br. eur. et E. dhersifoUo (Schleich) Br. eur. comparanda, ab hie nervo folii caulini ex- tremitate dorso haud denticulo instructo, foliis rami acutis et rotundato-obtuso, ab illa nervo folii ramei valido et extremitate dorso denticulo instructo, theca angustiore, opercule breve rostrato dignoscenda. Rhynchostegiuin ovalifolium Sh. Okamura. sp. nov. Habitat in terra irrorata. Plantac robustae, depresso-caes- pitosae, caespitibus densiusculis, viridibus, in statu sicco lutescenti- viridibus vel stramineois, sericeo-nitidiusculis, mollibus. Caulis repens, usque ad 4 cm. longus, densiuscule fuscescenti-radiculosus, densiuscule et irregulariter pinnatim ramosus, sectione teres c. 0.3-0.4 mm. diam., fascicule centrah c. 30-40 ,« magno, reti inter- medio hyaline, cellulis hexagonis teneUis c. 28-35 ,« magnis, celluhs periphericis c. 2-3-seriatis minoribus c. 9-12,« magnis incrassatis lütescentibus ; ramis prostratis, usque ad 3 cm. longis, simpücibus vel 1-3 ramulosis, dense et imbricato-foliosis complanatis, apice breviter attenuatis vel obtusis, cum foliis c. 3.5-4.0 mm. latis. Folia sicca paulum longitidinaliter rugosa, madida patentia, vix decurrentia, concava, haud plicata, e basi angusta late ovata in CONTKIBUTIONES NOV.\E AD FLORAM BRYOPHYTON JAPONICAM. 95 acumen subito clongatum subulatum semitortum attenuata, c. 2.1-2.4 mm. longa et c. 1.2-1.4 mm. lata, marginibus ubique seiTulatis ; nervo infra apicem folii evanido, in sectione transversal! plano-convexo, dorso valde prominenti, basi c. 50/^ lato et c. 40// crasso, e 4-stratis cellularum uniformium composite, ceUulis ven- tralibus c. 4-5 majoribus, dorsalibus c. 8 ; cellulis laminalibus laevissimis, valde cMorophyllosis, prosenchymaticis, c. 70-130 // longis et c. 7-8 /^ latis, basilaribus brevioribus laxioribus c. 50-60 j" longis et c. lO-lo r latis, in angulis concavis rectangulis numerosis (c. 20). Inflorescentia autoica ; flores masculi in foliorum axiUis ramorum ; folia perigonialia intima ovalia, apice acuta, integerrima, onervia, c. 0.6 mm. longa et c. 0.35 mm. lata ; antheridia c. 10 ; flores feminei in ramis. Ramulus perchaetlaUs valde radiculosus. Bracteae perichaetil intimae e basi alte vaginante sensim auguste subulatae, subula recurva minte serrulata vel subintegerima ; nervo tenui, medio evanido ; ceUulis basilaribus laxioribus. Seta c. 7-10 mm. longa, erecta, rubra, laevis, sicca torta. Theca inclinata, oblonga, asymmetrica, gibbosa, cum collo c. 1.5-1.6 mm. longa et c. 0.64-0.72 mm. crassa, fusca, laevis, collo obconico ; ceUulis exothecii hexagonis oblongo-hexagonis, ad orificium in seriebus duabus minutis ; sto- matibus in collo numerosis. Exostomii dentés lineari-lanceolati, luteo-fusci, hyaflne limbati, basi transversim striatuli, apice hyalini et minute papiUosi, intus dense et alte lamellosi ; endostomium luteum ; corona basilaris plicata, laevis ; processus dentium fere B. X15 Fig m 41 . Rhyncliostegium ovalifolium. A. riauta ( X 1) ; B. Folia ( x 15) ; C. Sporangia ( x 10). 96 ART. 4. — SHUTAI OKAMURA : longitudinis, carinati in carina anguste perforati, ininutissime papillosi ; cilia bina, bene evolnta, nodulosa, rainntissime papulosa. Sporae c. 14/>« magnae, virides, laeves. Caetera ignota. Hondo : Prov. Ise, Unomori-jinslia, Yoktaiclii-shi (Leg. Kititaro Mura- TA ! 28. Xn. 1914.). Statura faciesque omnino 7?. imllidifolio (Mitt.) Jaeg ; foliis autem ovalibiis apice brevioribus, seta breviore, theca minore jam abunde diversa. ühynchostegiuui spiralifoliiim Sh. Okamuba. sp. nov. Habtat in solio lacus c. 2 m. profunde. Plantae robustae, valde depresso-caespitosae, caespitibus late extentis, lutescenti- fuscescentibus, in statu sicco nitidiusculis, rigidiusculis. Caiilis repens c. 6-20 cm. longus, hie illic fasciculatim fusco-radiculosus, densius- cule et régularité!' pinnatim ramosus, sectione oblongus, c. 0.24- 0.28x0.16-0.2 mm. magnus, fascicule centrali nuUo, reti centrali Fii/, 42. — Rhyncliostegiuni spiralifolixun. A. rianta ( x 1) ; Ü. Folia ( x 20). CONTRIBUTIONES NOVAE AD FLORAM BRYOPHYTON JAPONICAM. 97 hyalino, cellulis hexagonis tenollis c. 15,« magnis cellulis pori- phericis c. 6-soriatis valdo iiicrassatis liitescentibus vol fuscis ; ramis prostratis, c. 2-2.5 cm. longis, strictis, simplicibus vol parco pinnatim ramulosis ; caulibiis ramisqiu^ donsiiisculo foliosis valclo complanatis, emu foliis 4-5 mm. latis, apico obtusis. Folia iit vidotm* disticha, sicca paulum riigulosa apice spiraliter torta,, madida patentia, hand decm'rentia, anguste oblongo-lanceolata, in acnmen elongatnm snbn- latnm hand tortnm sensim attennata, c. 2.4-2.8-3.4 mm. longa et c. 0.4-0.5-0.64 mm. lata, eoncavinscnla, marginibns ubiqno intoger- rimis ; nervo valido, modinm fohi evanido, lutoo, in sectione trans- versali piano -convexo, dorso prominenti, basi e. (30 /-« lati et c. 35 f« crasso, e stratis 3 ceUularnm nniforminm composito, cellnlis ven- trahbus c. 6, dorsalibus c. 10 ; cellnlis laminalibns laevissimis, pro- senchymaticis, c. 70-120,« longis et c. 7/^ latis. snperioribns brevi- oribns c. 80 /^ longis, basilaribns brevioribns, ad insertionem ob- longis 1-3-seriatim dispositis c. 40-50/^ longis et c. 12-14/^ latis parietibns incrassiusculis hand porosis, in anguUs hand concavis panels breviter rectangnlis vol qnadratis saepe inconspicnis. Caetera ignota. Hondo : Prov. Ködnke, in Lacn Marunnma [% ?S] (Leg. Haraikawa ! VU. 1914.) ; Prov. Bizeu, in Stagno in Tatsimokucliiyama, Mitsn-gun (Leg. Usuô Uno! 5. X. 1915.). Nomen speciei ab foliis in statu sicco apice spiraliter tortis. 98 ART. 4. SHÜTAI OK AM Uli A INDEX. PAGE. AërobryoïDsis assimüis (Caed.) Bkoth. . . 31 A. Parisii (Card.) Beoth 29 A. subdivergens (Broth.) Broth 31 Aërobrynm Ferriei Broth 29 Amplntüuin Lipponicum (Hedw.) Schimp. 10 Aulacomitrium minntissimum Sh. Oka- MUEA 16 Aiilacomniaceae 21 Aiüacomniiim targidiim (Wahlenb.) Sch- TVAEG 21 Barbiila Orientalis (Wils.) Beoth 8 Bartramiaceae 21 Bazzania s^jinosa Sh. Okamuea 2 Braclaytheciaceae 76 Bracliytbeoium phimosnm (Sw.) Bk. eue. 90 Bryhnia ussxiriensis Beoth 83 Bryaceae 18 Bryuni argenteum L 18 l/eratodon pnrpnreus (L.) Beid 5 Cbiloscypbus polyanthus (L.) Ooeda. var. rinilaris (Schrad.) Nees 2 C. riviilaris (Schead.) Loesk 2 Cümaciaceae 25 Climaciiim japonicum Lindb 25 Crypliaeaceae 25 Dicranaceae 5 Dicranella beteromalla (Dill , L ) Schimp. 5 Dicrannm Bergeri Blandow 5 D. ma jus Tuen 5 D. Schraderi Schwge 5 D. scoparinm (L.) Hedw 6 Dolicbomitra cymbifolia (Lindb.) Beoth. var. snbintegorrima Sh. Okamuea. . . 47 Drepanocladus nncinatns (Hedw.) Warnst. 59 Ectropotheciiim Sbiragae Sh. Okamuea. . . 60 Entodon arenosus Sh. Okamuea 48 E. dolicbùcncnllatiis Sh. Okamuea 49 Eiitodoiitaceae 48 PAGE. Eurrbyncbium Arbusciila Beoth 91 E. yezoanum Sh. Okamuea 91 Xabroniaceae 54 Fissidens japonicus Doz. et Molk 6 F. lateralioides Sh. Okamuea 6 F. nagasaldniTS Besch. 8 F. ziijpelianns Be. jav 8 Fissidentaceae li Frullania moniliata Nees 3 Funaria bygrometrica (L.) Sibth 18 Funariaceae 18 Glypbomitriiun angiistifolium Sh. Oka- muea 9 Grimmia Hisaucbii Sh. Okamuea 10 G. lüyosbii Sh. Okamuea 12 Grimmiaceae 9 Hajiloclatlium capillatum (MiTT.) Beoth.. 58 H. latifoUum (Lac.) Beoth 58 Hai)lobymemum bi£orm(! Beoth 54 Hedwigia albicans (Web.) Lindb 25 Hedwigiaceae 25 Herpetineuron attentiatus Sh. Okamuea. . . 54 Himantocladiiim loriforme (Be. jav.) Fleisch 37 Hookeria nipponensis (Besch.) Beoth. . . 54 Hookeriaceae 54 Hygrobypunm ocbracejim (Trun.) Beoth. 59 Hylocomium pyrenaicum (Speuc.) Lindb. 60 H. splendens (Hedw.) Schimp 60 Hypniim hakoniense Mitt 80 H. Scbreberi Wild 60 Hy2)naceae 59 Isbibaea japonica Iîeoth. et Sh. Oka- BIUEA 85 Jsopterygium expallescens Levieb 65 I. Giraldii (C. Müll) Pae 65 I. Hisaucbii Sh. Okamuea 65 I. taxirameum (Mitt.) Jaeg 67 CONTllIBUTIONES NOV^VE AD FLOlîAM 13RY0PHYT0N JAPONICAM. 99 PAGE. J imgormaniacoiio iikrogyao 2 Xjembopliyllîxceae 47 Leskea pusilla ^Mitt 56 Leskoacc^ae 51 Leiicodon perdeijendeus Sh. Okamura. . . 25 Loiicodontaceae 25 Macromitrium incuniim (Lindb.) Palis. 18 MtKlotlieca Perottetiana Mont 3 Marchantia tosana Stkph 1 Marchantiiiceae 1 Mcteoriella cnsjsidata Sh. Okajiuea. . . 34, 37 M. Kutôi Sh. Okamuea 36, 37 M. soluta (Mitt.) Sh. Okamuea. . . 35, 37 Meteoriojisis aucistroides (Hen. et Caed.) Beoth 31 IMeteorium ancistroides Ben. et Card. . , 31 31. liymalayense Par , . . . 31 M. Parisii Caed . 29 M. solutiim Mitt 35 ]\Iniaceae 19 Mniiim Kiyoshii Sh. Okamuea 19 M. ligulifolium Caed 20 M. piinctatum (L.) Hedw 20 M. subglobosum Be. eue 20 M. trichomanes Mitt 21 Myuroclada concinua (Wils.) Besch. . . 90 rfeckera idumoana Sh. Okamuea -40 N. kamakurana Sh. Qkamuea 42 Neckeraceae 27 Neckeropsis j^seudonitidula Sh. Okamuea. 39 Okamiiraea brevipes Beoth 76, 85 O. cristata Beoth 78, 85 O. cristata Beoth. var. gracilis Beoth. 79, 85 O. cristata Beoth. var. miütiflagelliferoa Sh. Okamuea 85 O. hakoniensis (Mitt. ) Beoth 80, 85 O. imbricata Beoth 82, 85 O. plicata Caed 85 O. iissuriensis (Beoth.) Beoth 83, 85 Orthotrichaceae 16 Oxyrrhynchium Schottmiilleri (Beoth.) Beoth 90 O. ScliothomüUeri (Beoth.) Beoth. var. perlougicladiim Sh. Okamuea 90 Philonotis palustris Mitt 21 Physconiitrium curystomum (Nees.) Sendt. 18 P. subeurystomtim Caed 18 Pilotrichopsis dentata (Mitt.) Besch. . . 25 page. l^innatera formosana Sh. Okamuea 45 Plagiocbila f rnticosa Mitt 2 I'leuroziopsis rutbenica (Weinm.) Kcndb. . . 25 Pogonatum arisanense Sh. Okamuea. . . 21 P. grandi folium (Lindb.) Jage 24 1'. iuflexnm Lindb 24 P. uriiigerum Schimp 24 Polytricbaceae 21 Polytricbum alpimun L 24 P. commune L 24 Pottiaceae 8 Pterobryopsis cucuUatifolia Su. Okamuka. 27 Ptychodium jierattenTiatum Sh. Okamuea. 87 Pylaisia laeto-viridis Sh. Okamuea."... .. 51 Radula complanata (L.) Dum 3 Beboulia baemispbaerica (L.) Raddi. . , 1 Ebacomitrium canescens (Weis., Timm ) Beid 13 11. fasciculare (Schead.) Beid 13 11. bypnoides (L.) Lindb 13 11. Iwasaldi Sh. Okamuea 13 llbacopilaceae 54 Bbacopilum aristatum Mitt 54 Ilba]ibidostegium argutum Sh. Okamuea. 72 11. dcmissum (Wils.) De Not 74 R. japonicum Beoth 74 Kbizogoniaceae 21 Ebizogonium spiniforme (L.) Beuch. . . 21 Ebodobryum Wicburae (Beoth.) Beoth. . . 18 Ebyncbostegium ovalifobum Sh. Okamuea. 94 E. si)iralifoUum Sh. Okamuea 96 Eliytidiadelpbus calvescens (Wils.) Beoth. 59 E. squarrosus (L.) Waenst 59 E. triquetrum (L.) Waenst 59 Ebytidium rugosum (Eheh.) Ivindb. . . 59 Scbwetsobkea Matsumurae Besch 54 SematopbyUaceae 72 Spbagnaceae 4 Spbagntim acutifolium Eheh 4 S. Girgensobnii Euss.; 4 S. japonicum Warnst 4 S. pajiillosum Lindb 4 S. sqimrrosum Pees 4 Stereodon cymbifobus Sh. Okajmuea. . . 62 S. fertilis (Sendtn.) Lindb 65 S. plumaeformis (Wils.) ÄIitt 65 Tbamnium Fauriei Beoth. et Pak 47 Tbuidium japonicum Doz. et Molk. . . 58 100 ART. 4. SHUTAI OKAMURA Tortiüa emarginatfi (Doz. et JIolk.) Mitt. 8 Trachypns linmilis Lindb 37 Trematodon drepanellns Besch 5 Ï. longicollis Michx 5 Trichocolea tomentella (Htjds.) IiIndb. V esiciilaria Sasaokae Sh. Okajiuba. . . V. Sliiinadao Sh Okamuka iVeisia Airidiila (L.) Hedw PAGE. . 3 . G7 . G9 8 Published July 25th, 1916. JOUENAL OF THE COLLEGE OF SCIENCE, TOKYO IMPERIAL UNIVERSITY. VOL. XXXVIII., ARTICLE 5. On some Japanese Calcareous Sponges belonging to the Family Heteropiidse. By SaDJi Hözawa, RigakusM. Zoological Institute, Science CoUege, Tokyo Imperial Universit}'. WUli 2 plates. In the collection of calcareous sponges collected by Professor Ijevia and preserved in the Zoological Institute of the Science College, there exist seven species belonging to the family Hetero- piidge. Six of them seem to be new to science. In the present paper I propose to give descriptions of aU the species in the fol- lowing order : 1. Grantessa shimeji, n. sp. 2. „ sagamiana, n. sp. 3. „ intusarticuJata (Cakter). 4. „ hasipapillata, n. sp. 5. „ mitsukurii, n. sp. 0. Heteropia striata, n. sp. 7. Amphiute ijimai, n. sp. Here let be fulfilled my pleasant duty of expressing my hearty thanks to Professor Ijima, who has not only kindly placed at my disposal the valuable fruits of his long years' collecting, but has 2 ART. 5. — SANJI hözawa: also rendered me many a courteous help during the course of my investif^ation. Family HETEROPIIDiE Dendy. Genus Grantessa von Lendenfeld (emend.). 1. Grantessa sJiimejif n. sp. (PL I., figs. 1, 2 ; PI. n., figs. 10, 11 ; textfig. 1). This new species is represented in the collection by five speci- mens. They differ more or less from one another in certain minor points of external feature, but are practically identical in the finer structure. The first specimen (Sei. Coll. Spec. No. 1 ; PI. I., figs. 1, 2), which is taken for the type, was obtained by K. Aoki in the immediate neighbourhood of the Misaki Marine Biological Station. The sponge forms an irregular hemispherical colony with a height of 45 mm. and maximum diameter of about 85 mm. In the superficial parts, the colony is seen to consist of numerous sub- cylindrical, erect and on the whole radially directed tubes, each of which somewhat tapers distally and terminates with an osculum. The tubes may be 6-10 mm. long, and IJ mm. broad in the lower parts, where the wall presents a thickness of 0.35- 0.6 mm. The wall gradually diminishes in thickness towards the oscular margin. The osculum is approximately circular with a diameter of J-1 J mm. An oscular fringe of oxea is present, but is scarcely visible to the naked eye. The dermal surface is rough, owing to the presence of oxea in scattered tufts projecting almost vertically fi'om it. The gastral surface is also rough on account ON SOME JAPANESE CALCAREOUS SPONGES. 3 of the projecting apical rays of gastral qiiadriracliates. Several of the tubes are seen to bear diverticuluna of varying length, the blind end of which is outwardly directed and may sometimes show a perforation representing osculum at an early stage of breaking through. The deeper parts of the colony give quite a different view from the peripheral. The inner continuations of the radial tubes branch and anastomose with one another and thus form a complicated network of tubes which constitutes the greater part of the mass of the colony. Here the diameter of the tubes and the thickness of wall are somewhat greater than in more peripheral parts of the colony. The sponge is colourless in alcoliol, its texture rather firm and compact. The second specimen (Sei. Coll. Spec. No. 2) was also obtained by K. AoKi at the same locahty. It is an irregularly shaped mass, 80 mm. high, and 40 mm. in maximum diameter. The os- cular tubes are on the whole of a larger calibre than those of the type- specimen. They reach 14 mm. in length and 5 mm. in width in the widest part, where the wall is nearly J mm. thick. The dermal surface is relatively smooth, the tufts of oxea being not so strongly developed as in the type- specimen. The third specimen (Sei. Coll. Spec. No. 3) was obtained by Prof. Ijima at the same locahty. It is a small colony, provided with about 30 oscular tubes. In external characters it agrees fairly well with the type -specimen. The fourth specimen (Sei. Coll. Spec. No. 4) was collected by Mr. K. Yendô at Öshima, in the Province of Shima. It is a little smaller than the type- specimen. Height, 30 mm. Maximum dia- meter, 65 mm. Minimum diameter, 40 mm. The fifth specimen (Sei. CoU. Spec. No. 5) hails from the same locality as the preceding. It represents an irregularly 4 ART. 5. — SANJI HÖZAWA : roundish mass, measui'ing about 15 mm. in height and 30 mm. in maximum diameter. On the whole, the oscular tubes are much smaller than those of any other specimen. The largest tube meas- ures only 5 mm. long and 2 mm. broad in the widest part, where the wall is nearly 0.45 mm. thick. The osculum is irregularly circular, measuring about 0.45 mm. in diameter ; oscular fringe of oxea not visible with the naked eye. The following description of canal system and spicules is "based on studies of the type- specimen. Canal system (PI. n., figs. 10, II). The canal system is typically syconoid. Dermal pores, 40- 100 fji in diameter, are scattered irregularly over the sponge surface. The dermal cortex is rather weakly developed with a small quantity of mesoglœa. Tlie flagellate chambers are straight, cyhndrical and usually not branched ; they are radially arranged around the gastral cavity, extending from gastral to dermal cortex. Each chamber is provided with a number of small prosopyles of 10-40 /'- diameter, and with a single apopyle at the inner end. The position of nucleus in collar cells is apical. The gastral cortex is nearly as thick as the dermal and is perforated by short but rather wide exhalant canals, which usually arise each from a single chamber and sometimes from two or three chambers. A weakly developed diaphragm occurs at the apopyle. Diameter of exhalant pores 70-150 /a Skeleton (PL II., figs. 10, 11). ' The dermal skeleton is composed of the following elements : 1) triradiates, which are tangentially arranged in several layers without definite orientation ; 2) the paired rays of subdermal ON SOME JAPANESE CiVLCAREOUS SPONGES. 5 pseudosagittal triradiates ; and 3) oxea which project nearly verti- cally from dermal surface and which are grouped in small tufts standing in no definite relation to radial chambers. The tubar skeleton near the osculum is of the inarticulate type, consisting of the basal rays of subdermal pseudosagittal and subgastral sagittal triradiates (PL II., fig. 11). In the remaining parts of the sponge, where the wall is thicker and the chambers are more elongate, the tul)ar skeleton receives an addition of some rows of sagittal triradiates with outwardly directed basal rays and is thus of the articulate type (PL II., fig. 10). The gastral skeleton is made up of tangentially placed tri- radiates, of the paired rays of subgastral triradiates and of the facial rays of gastral quadriradiates, of which the apical rays pro- ject into the gastral cavity. The skeleton of the oscular margin is composed of oxea and triradiates. The oxea I'lm longitudinally and parallel with one îinother as well as with the downwardly directed basal rays of the triradiates. Spicules. Dermal triradiates (textfig. 1, a, h). — Slightly sagittal. Basal ray smooth, straight, ending in a sharp point, about 80 /^ long and 8 !'- thick. Paired rays longer and slightly thicker than basal ray, not quite smooth, almost straight but sometimes slightly crooked, gradually tapering and sharply pointed, about 120 ,« long and 10 /^ thick. Subdermal triradiates (textfig. 1, c, d, e). — Pseudosagittal. AU rays equally thick, lying nearly in the same plane. Basal ray longer than paired rays, straight excepting slight curvature near base, sometimes very slightly crooked, gradually tapering, sharp- ART. 5. — SANJI HÔZA"WA Textfig. 1. Grantessa sMmeji. a, b, Dermal triradiates. c, cl, e, Subdermal triradiates. f, g, Tubar triradiates. li, i, j, Subgastral triradiates. k, 1, Gastral triradiates. m, n, Gastral qnadriratliates. o, p, Triradiates of oscular margin, q, r, s, Oxea of oscular margin. (All figs. 200 X ). pointed, 120-180 <« long and 10 z^« thick. Paired rays unequal in length and shape ; the longer ray sometimes nearly as long as the shorter, curved, 90-140/^ long and 10 /< thick. The shorter ray almost straight, 60-110/^ long and 10/-« thick. Tubar triradiates (textfig. 1,/. f/).— Sagittal. Basal ray long, almost straight, gradually tapering and sharply pointed, 120-270 /^ ON SOME JAPANESE CALCAREOUS SPONGES. 7 long and 10 /^ thick. Paired rays nearly equal, curved, gradually tapering and sharply pointed, 70-100,« long and 10/^ thick. Subgastral triradiates (textûg. 1, h,i,j). — Sagittal. Basal ray long, straight, gradually tapering, sharp-pointed, about 250 ,« long and 10 // thick. Paired rays nearly equal, widely diverging, lying not in the same plane as the basal ray, shghtly curved, gradually tapering to a point, about 80 /^ long and 12 ,« thick. Gastral triradiates (textfig. 1, k, I). — Sagittal. Basal ray much longer than paired rays, quite straight, tapering from base to the sharp point, 140-270,« long and 8-10/^ thick. Paired rays straight, almost equal or slightly differentiated in length, some- what thicker than the basal ray, gradually tapering, sharply pointed, 90-120 /^ long and 10-12 /^- thick. Gastral quadriradiates (textfig. 1, m, n). — Almost like gastral triradiates with addition of a short apical ray. Basal ray much longer than paired rays, quite straight, gradually tapering, sharp- pointed, 150-200/^ long and 8-10/^ thick. Paired rays almost straight, subequal in length, slightly thicker than the basal ray, gradually tapering, sharply pointed, 100-120/^ long and 10-12/^ thick. Apical ray much shorter and slender than facial rays, ending sharply, slightly curved and directed upwards, 30-50 /^ long and 8-10 [j- thick. The quadriradiates situated near the osculum have the paired rays curved and more widely diverging than in those of other parts. Triradiates of oscular margin (textfig. 1, o, p). — ^Basal ray smooth and straight, gradually tapering, sharp-pointed, slightly longer than paired rays, 100-150 /^ and 10 /^ thick. Paired rays stouter than the basal, widely diverging, almost at right angles to the basal ray, curved, not quite smooth, sharply or bluntly pointed at end, 100-130/^ long and 12/^ thick. 8 ART. 5. — SANJI hüzawa: Oxea. — Usually slightly curved, nearly uniformly thick through- out their length and sharply pointed at both ends, of which the free end is sometimes provided with a more or less distinct nodi- form ring. They occur in tufts projecting from the dermal surface and with the inner J-J of their length imbedded in the chamber layer. Oxea of oscular margin (textfig. 1, q, r, s). — Similar to those just described, but on the whole longer, being 100-320/-« long and 6-8 /^ thick. Note. — The above described species can not be identified with any previously described. The specific name " Shimeji " is given to it on account of its resemblance in form to certain fungus known in Japan by that name. This sponge is common in the neighbourhood of the Misaki Marine Biological Station. It occiu-s in especial in abundance in the spring, attached on perpendicular or overhanging faces of rocks below low tide- mark. Localities. — Misaki ; Öshima, Province of Shima. 2. arantesaa saganiiana, u, sp. (PI. I., fig. 3 ; PI. n., fig. 12 ; textfig. 2). The type-specimen of this new species (Sei. Coll. Spec. No. 20) was collected from a depth of 429-572 m. at Okinose, Sagami Sea. Sponge consists of a sohtary person ; elongate, cylindrical, slightly laterally compressed and bent at base, with an indistinct oscular fringe ; surface slightly hispid, due to large oxea and trich- oxea projecting here and there. The body is about 45 mm. long and 9 mm. in gi-eatest breadth. It narrows towards both ends. The sponge wall is about 1 mm. thick in the middle parts of body ; nearer the osculum the thinner it becomes. The osculum ON SOME JArANESE CALCAREOUS SPONGES. 9 is elliptical, G mm. by 3 mm. wide, and leads into the wide and deep gastral cavity. The colour in alcohol is greyish white ; the texture is fairly firm and elastic. A second specimen (Sei. Coll. Spec. No. 21) was obtained l)y Prof. Ijima in the Saganii Sea, off Cape Sunosaki, from a depth of 429 m. It was attached to the spine of Goniocidaris mikado (Döderlein). The specimen is much smaller than the type, meas- uring 22 mm. in length, 5 mm. in greatest breadth and about 0.8 nnn. in thickness of wall. The surface is more strongly hispid than in the type. The osculum is elliptical, oj mm. by 2 mm. wide ; it is in part provided with a distinctly developed oscular fringe. The third specimen (Sei. Coll. Spec. No. 22 ; PL 1., fig. 3) was obtained by K. Aoki at the entrance to Enoura, Suruga Bay, from a, depth of 380 m. It was found attached at base on a coral together with a polyzoan colony. It represents an elongate, slightly curved and laterally compressed tube, with the surface thickly beset with projecting oxea. Total length 30 mm. Greatest breadth 7 mm. Wall not thicker than 1 mm. The osculum is circular with a, diameter of 3 mm. Oscular fringe is rather distinct. The following account refers to the type-specimen. Canal system (PI. n., fig. 12). The canal system is sy conoid. Dermal pores small, circular or oval with a diameter of 40-150 /^, thickly distributed aU over the surface of sponge ; the pores lead singly, or more frequently several together in groups, into wide inhalant canals through the thin dermal cortex. The inhalant canals extend inwards, around and between the distal ends of flagellate chambers. Flagellate chambers are arranged radially with regard to the 10 ART. 5. — SAN.JI HüZAWA : central gastral cavity, extending nearly through the entire thick- ness of the sponge wall. They are broad in the proximal parts and become markedly narrower in the distal parts. Sometimes they are seen to be united with one another in the proximal parts and also to divide distally into branches. The nucleus of collar cells occupies an apical position. The gastral cortex is very thin with a small quantity of mesoglœa ; it is pierced by very short exhalant canals, arising from either a single or fi-om 2-4 flagellated chambers. A well- developed diaphragm exists at the apopyle. The exhalant canals open into the gastral cavity through small angular meshes, meas- uring 200-400 /^ across and formed by intercrossing rays of gastral tri- and quadriradiates. Skeleton (PL n., fig. 12). The dermal skeleton is made up of : 1) triradiates tangential - ly disposed in a few layers, with basal ray generally directed downwards ; 2) the paired rays of subdermal pseudosagittal tri- radiates ; 3) large oxea, which occur here and there in vertical disposition in the sponge wall and which usually project outwards on the dermal side only but may sometimes do so also on the gastral side ; 4) trichoxea which are but rarely found. The tubar skeleton is composed of : 1) the centripetal basal rays of subdermal pseudosagittal triradiates ; 2) the centrifugal basal rays of subgastral triradiates ; and 3) one or two intermediate rows of triradiates which, have their basal ray directed outwards. The above refers to the thick-walled middle parts of the sponge. In the parts close to osculum where the wall is thin, the tubar skeleton consists only of the basal rays of subdermal and sub- gastral triradiates. ON SOME JAPANESE CALCAREOUS SPONGES. 11 The gîisfcral skeleton consists of the paired rays of subgastral sagittal triradiates, of triradiatos, and of the facial rays of quadri- radiates. Of the last, the short apical rays project into the gastral cavity. Triradiates and quadriradiates are disposed tangentially, otherwise in no definite orientation. The former are always less numerous than the latter. The skeleton of the oscular margin consists in a close inter- lacement of fine longitudinal trichoxea and of triradiates and quadri- radiates, both which have strongly divergent paired rays and downwardly directed basal ray. To the above there may be added a number of large oxea which run longitudinally and parallel with trichoxea. Spicules. Dermal triradiates (textfig. 2, (t, h). — Slightly sagittal. All rays straight, smooth and gradually sharp-pointed. Basal ray slightly longer than paired rays, equally thick or a little thicker than these, 240-370 /^ long and 20-28 /^ thick. Paired rays of nearly equal length, 200-270 p- long and 20 ,« thick. Subdermal triradiates (textfig. 2, c, d, c). — Pseudosagittal, ir- regular. All rays of nearly same thickness but of different length and shape. Basal ray much longer than paired rays, sharp-pointed, its basal parts slightly curved and for the rest straight or nearly so, 200-490/^ long and 16-20/^ tliick. Paired rays are of diff'erent length and shape, lying not in the same plane with basal ray. The longer of them gradually tapers to a sharp point, is bent near base and is sometimes more or less crooked in the remaining parts, 100-160/^- long and 12-16/^ thick. The shorter ray is sharp- pointed and strongly curved in the middle parts, 90-110/^- long and 16-20,« thick. 12 ART. 5. SANJl HOZAWA : Textfig. 2. Graniessa sagamiana. d, e, g. i, k, ni, P. r, s. Dermal tiiradiates. Subdermal triradiates. Tubar triradiates. Siibgastral triradiates. G astral triradiates. Gastral qiiadriradiates Triradiate of oscular margin. Qnadrira'- thick. Gastral triradiates (textfig. 2,/, /.). — Slightly sagittal. All rays equally thick. Basal ray quite straiglit, smooth, tapering from base to the sharp point, 180-270/^ long and l()-20 ,« thick. Paired rays equal in length, almost straight excepting a slight curvature near base, gradually sharp-pointed, 170-210 « long and 10 ,« thick. Gastral quadrir a diätes (textfig. 2, /, m). — Facial rays exactly similar to gastral triradiates. Basal ray 170-280//- long and 11]- 20," thick. Paired rays 200-210// long and 10-20// thick. Api- cal ray curved, sharp-pointed, nearly as thick as the facial rays but much shorter, 50-70// long 10// thick. Triradiates of oscukir margin (textfig. 2, n). — Basal ray quite straight, sharply pointed, 140-280 /^ long and 8-12 // thick. Paired rays slightly curved, standing nearly at right angles to basal ray, thicker and shorter than the latter, 100-170 // long and 12-lQ/Jt thick. Quadriradiates of oscular margin (textfig. 2, o, ^j). — Facial rays exactly similar to triradiates of the oscular margin. Apical ray very short, slightly curved. Basal ray 200-350 /^ long and 12// thick. Paired rays 150-220// long and 12-10// thick. Large oxea (textfig. 2, q, r, s). — More or less curved, of vary- ing lengths, nearly uniformly thick in the greater part of their length but tapering at both ends which are sharply pointed, 0.0- 1.8 mm. long and 30-50 // thick. 14 ART. 5. — SA^^^ hözawa : Trichoxea of oscular margin. — Slender, quite straight, sharply pointed at inner end, generally broken off at outer end. A large example with broken outer end measured 590 /^ in length and 4 // in thickness. Trichoxea of dermal cortex. — Slender, hair-like, generally much thinner than those of the oscular margin, measuring about 2 /^ in thickness. Note. — This form seems to be quite distinct from any of the hitherto known species. Grantcssa lanceolata (Beeitfuss)^^ may be looked upon as its nearest ally, though showing some marked differences in spiculation. Localities. — Okinose ; off Sunosaki ; Entrance of Enoura, Su- ruga Bay. 3. Grantessa intusavticulata (Carter). (PI. I., figs. 4, 5 ; PL n., fig. 13 ; testfig. 3). Uypograntia intusarticulata, Cakter, 1885-1886 (1), p. 45. „ medioarticulata, Cabtee, 1885-1886 (1), p. 46. Grantessa intusarticulata, Dkndy, 1892 (1), p. 108 ; 1893 (2), pp. 181, 201, PI. Xm., fig. 18, Grantia intusarticulata, Breitfuss, 1897 (2), p. 219. Sixteen specimens of this species have come under my ob- servation. Thirteen of them (Sei. Coll. Spec. Nos. 6-17, 29) were collected by K. Aoki at Doketsba in the Sagami Sea from depths varying from 185 to 214 meters ; one (Sei. Coll. Spec. No. 30) came from Jôgashima (Misaki), while the remaining two (Sei. Coll. Spec. Nos. 31, 32) were obtained by Prof. Ijima from the shallow bottom in the neighbourhood of the Misaki Marine Biological Station. All specimens in the first group are sohtary tubular indivi- duals attatched by the narrowed base. They are provided with an 1). Ebnerella lanceolata Bbeitfuss, 1893 (3), p. 28, Taf. I., Fig. 3-5 ; Tuf. IV., Fig. 24, 25. ON SOME JAPANESE CALCAREOUS SPONGES. 15 oval or circular osciilum, surromided by a more or less distinctly developed fringe of oxea. The specimen from Jôgashima is a fragment of oscular tube with base. The last two specimens re- present a colony of several small and tubular individuals joined together at their base (PL I., fig. 5), To base further description on I have selected one of the Döketsba specimens (Sei. Coll. Spec. No. 6 ; PI. I., fig. 4). It mea- sures about GO mm. in total length and 11mm. in greatest width. The wall reaches about 1 mm. in thickness. The osculum is oval, measuring 2J by 5 mm. It leads into a wide and deep gastral cavity. The sponge is bent in the basal parts, whe^e it presents a somewhat irregular contour and also gives rise to some tubercular processes for attachment. The dermal surface looks very smooth, while the gastral is somewhat rough owing to projecting apical rays of gastral quadriradiates. Colour in alcohol greyish white. Texture firm, but elastic. Canal system (PI. n., fig. 13). The canal system is typically syconoid. The inhalant canals open on the dermal surface by means of small irregularly round- ish, mesh-like pores of 70-100 ,« diameter. The canals, after tra- versing the dermal cortex, unite into larger trunks which lead into the interstices between flagellate chambers. The dermal cortex, which appears finely hispid on account of vertically disposed micr- oxea, is rather thin, being about 100 // thick ; it directly overhes the distal ends of flagellate chambers. The flagellate chambers are elongate and radially arranged in the chamber layer. They are usually simple, but are sometimes divided into two or three parallel and distally narrowing branches. The chambers communi- cate either singly or several together with exhalant canals, which 16 AET, 5.' — SANJI HÖZAWA : are short but relatively wide. Diaphragm is present at each apopyle. The nucleus of collar cells occupies an apical position. The gastral cortex is almost as thick as the dermal ; it is perforated in a mesh-hke manner by irregularly quadrate, penta- gonal or hexagonal openings (150-250// across) of exhalant canals. The epithehum lining the gastral cavity and exhalant canals is very conspicuous, the component cells being relatively large. The apical rays of gastral quadriracliates, as they stand out into the gastral cavity, seem to be covered all over by the epithelium. Skeleton (PL II., fig. 13). The dermal skeleton consits of triradiates, mieroxea and the paired rays of subdermal pseudosagittal triradiates. The triradiates lie parallel to the dermal surface with the basal rays in most cases directed towards the sponge base. The mieroxea are very small, numerous and thickly set. They lie in the dermal cortex at varying angles to the external surface, beyond which the outer ends freely project to a certain extent. In the middle parts of the sponge the tubar skeleton is formed by the centripetal basal rays of subdermal pseudosagittal triradiates, by the centrifugal basal rays of subgastral sagittal triradiates, and by several intermediate rows of sagittal triradiates, of which the basal rays are centrifugally directed. The gastral skeleton is chiefly made up of triradiates tan- gontially placed without definite orientation and arranged in several layers. It also contains the paired rays of subgastral triradiates as well as large quadriradiates, the apical rays of which project into the gastral cavity, pointing towards the osculum. The oscular margin is composed of trichoxea of varying thick- ness running longitudinally and parallel with one another, and of ON SOME JAPANESE CALCAREOUS SPONGES. 17 very closely set triradiates, which have very strongly diverging paired rays. Spicules. Dermal triradiates (textfig. 3, a, b). — ^SKghtly sagittal. Basal ray straight, smooth, gradually and sharply pointed, 120-250/^ long and 16-20/^ thick. Paired rays very slightly curved and gradually sharp-pointed. They are nearly as thick as the basal ray, but shorter, being 80-190/^ long and 12-16/^ thick. Subdermal triradiates (textfig. 3, c, d). — Pseudosagittal, irregu- lar. Basal ray considerably longer than the paired rays, shghtly bent near base, and gradually sharp-pointed, 280-360 ,« long and 16-20 ft thick. Paired rays equally thick but differing in length and shape. The longer ray gently curved and gradually sharp- pointed, 80-130 /i long and 12-16 /^ thick. The shorter ray more or less angularly curved in the middle, gradually and sharply pointed, 50-120// long and 12-16// thick. Tubar triradiates (textfig. 3, « long and 12 /^ thick. Gastral quadriradiates (textfig. 3, m, n), — Large and very stout, with gradually and sharply pointed facial rays which are equal or slightly differentiated sagittally, and with very strongly developed apical ray, curved and pointed only at the end. In ON SOME JAPANESE CALCAREOUS SPONGES. 19 typical cases, the facial rays are 180-230/^ long and 24/^ thick; the apical ray about 280 ,« long and 24 a« thick. Triradiates of oscular margin (textfig. 3, o, p). — Very strongly Siigittal. Basal ray longer and more slender than the paired rays, straight, gradually and sharply pointed, 230-310// long and 8 /^ thick. Paired rays strongly diverging, slightly curved, either gradually sharp-pointed or broadened in the middle and narrowed towards both the base and the pointed end. The oral angles are rather variable. Microxea (textfig. 3, q, r). — Nearly straight, symmetrically sharp-pointed at both ends, sometimes with a hastate point at one end, 76-92 ," long and 4 ,« thick, Trichoxea of oscular margin (textfig. 3, s). — Very slender, straight, 210-550 // long and 2 /^ thick. Oxea of oscular margin (textfig. 3, t, ii, v). — These resemble trichoxea, but are thicker, straight, almost uniformly thick through- out their length; the ends sharply pointed, 110-450 /^ long and 6 ft thick. Note. — The specimens seem to agree very well in all essential characters with the Australian species first described by Carter mider the names of Hypograntia intusarticulata and H. medioarticu- lata and later referred to Grantessa by Dendy. I am therefore strongly incHned to identify the Japanese form with that species. Localities. — Near Port Philip Heads (Carter) ; Watson's Bay, Port Jackson (Dendy) ; Dôketsba and Misaki, Sagami Sea. 4. Grantessa hasipaiyillata, u. sp. (PI. I., fig. 6 ; PL n., fig. 14 ; textfig. 4). This species is based on a single specimen (Sei. Coll. Spec. 20 ART. 5. — SANJI HÔZAWA : No. 23), which was obtained by K. Aoki at Döketsba, Sagami Sea. It represents in the main an irregnlarly cyhndrical sycon person ]:>roken off at one end and wliich has budded out near the damaged end a second, much smaller person. The mother person is slightly laterally compressed and exhibits several small nipple- shaped protuberances on the sides. It measures about 80 mm. long by 10 mm. broad at the widest part, where the wall is about 1 mm. thick. The osculum is circular with a diameter of 3J mm ; its margin is very thin and plain without fringe. The second person is cylindrical, circular in cross section, about 20 mm. long, 5 mm. broad, and 0.7 mm. in thickness of wall. It shows at the end a small osculum of only J mm. diameter. The dermal surface is quite smooth without any projecting spicules. The gastral sm*- face is lined with quadriradiates, of which the short apical rays project inwards. The gastral cavity is very wide and extends through the entire length of the specimen. The colour of the sponge in alcohol is greyish white. Texture is firm but ratlier brittle. Canal system (PI. n., fig. 14). The canal system is typically sy conoid. The dermal cortex is fairly thick. The dermal pores, measuring 40-100 /^ across, are thickly distributed over the dermal surface. They lead, either singly or two or three together, into prolonged inhalant canals extending deep into the chamber layer. The flagellate chambers are arranged radially with regularity. They are nearly straight, circular in transverse section, distally more or less narrowed, and usually, though not always, undivided. Their distal ends are aU in about the same plane, just beneath the dermal cortex. The position of nucleus in collar ceUs is apical. ON SOME JAPANESE C.\LCAREOUS SPONGES. 21 The gastral cortex is thinner than the dermal and the ex- halant canals are short. These spring each from a single or more flagellate chambers. Diaphragm present at apopyle. Gastral openings of exhalant canals are angular meshdike gaps, measuring about 0.2 mm. across on an average. Skeleton (PI. TL., fig. 14). The dermal skeleton is fairly well-developed, being composed of densely intercrossing rays of tangentially placed sagittal tri- radiates and of the paired rays of subdermal pseudosagittal tri- radiates. There is little regularity in the orientation of the dermal triradiates, save in the tendency of their basal rays to take ab- oscular direction. The tubar skeleton is typically inarticulate, being composed of the centripetal basal rays of subdermal pseudosagittal triradiates and of the centrifugal basal rays of subgastral sagittal triradiates. The gastral skeleton is more weakly developed than the ilermal. It is made up of slender triradiates and quadriradiates, l)oth tangentially placed but otherwise without definite orientation, and of the strongly developed paired rays of subgastral sagittal triradiates. Around the osculum are small oxea which are placed longi- tudinally and parallel with one another. Spicules. Dermal triradiates (textfig. 4, a, b, c). — Slightly sagittal. Basal ray is, as usual, shghtly longer and thicker than paired rays ; quite straight, tapering from base gradually to a sharp point, 130-460 /^ long and 20-48 /^ thick. Paired rays nearly equally long, slightly 22 ART. 5. — SANJI HÔZAWA ; Textfig. 4. Grantessa hasijxij)illata. a, b, c. Dermal triradiates. d, e, Rubdermal triradiates. f, g, Subgastral triradiates. h, i, j, Gastral triradiates. k, 1, Gastral quadriradiates. m, n, Oxea of oscular margin. (a— c, 66 X ; d— n, 100 x ). curved at base and tapering throughout their entire length, 150- 400 r- long and 16-36 ;« thick. Subdermal triradiates (textfig. 4, d, e). — Pseudosagittal, irregular. All rays are of nearly same thickness but of different length and shape. Basal ray nearly straight, sometimes slightly crooked, with sharp point, 360-480/^ long and 20-28/^ thick. The longer paired ray slightly curved and crooked, gradually tapering, sharp-pointed, 140-200 /^ long and 16-24 /^ thick. The shorter paired ray more straight than the longer, gradually and sharply pointed, 130-170/« long and 20-28 /^ thick. Subgastral triradiates (textfig. 4,/, g). — Strongly sagittal, strong- ly developed. Basal ray quite straight, tapering fi-om base to a sharp point, slightly thicker than the paired rays, 400-550/^ long ON SOME JAPANESE CALCAREOUS SPONGES. 23 and 28-40 /^ thick. Paired rays strongly diverging, almost of equal length, gradually tapering, sharply pointed, curved, lying in a plane different fi'om that of basal ray, 140-300 ^ long and 24-30 /^ thick. Gastral triradiates (textfig. 4, h, i,j). — All rays are of nearly equal thickness. Basal ray quite straight, tapering throughout their entire length, sharply pointed, 140-280 /^ long and 12-16 /^ thick. Paired rays nearly equal in length, either straight or slightly curved, gradually and sharply pointed, 130-200 ii long and 12-16 /i thick. Gastral quadriradiates (textfig. 4, Ä-, /). — Facial rays are exactly similar to gastral triradiates. Apical ray poorly developed. Basal ray 90-250/^ long and 12-16 i" thick. Paired rays 70-200,« long and 8-12 /^ thick. Apical ray smooth, slightly curved, sharply pointed, 20-30 ,« long and 12 i« thick. Oxea of oscular margin (textfig. 4, m, n). — Quite straight, spindle-shaped, usually thickest nearer proximal than distal end, tapering towards both sharply pointed ends. The fi'ee projecting ends usually broken off. Some complete examples measured 110- 240 /^ long and 3-4 /^ thick. Locahty. — Döketsba, Sagami Sea. 5. Grantessa niitsukurii, u. sp. (PI. I., fig. 7 ; PI. n., fig. 15 ; textfig. 5). The single specimen (Sei. Coll. Spec. No. 24), representing this new species, is a small colony of irregularly anastomosing tubes, the individuality of which is indicated only by the oscula num- bering six in all. Each osculum is surrounded by a very feebly developed fringe of oxea. The specimen was collected by the late 24 AET. 5. — SANJI HÖZAWA : Prof. MiTSUKüRi at Koajiro, close to the Misaki Marine Biological Station. The tubes measure 3-5 mm. in breadth, and the wall 1 mm. in thickness. The size of oscula is various, ranging from J-2 mm. in diameter. They lead into the common gastral cavity of a habitus corresponding to that of the entire specimen. The colour in alcohol is greyish white. The texture is rigid. Canal system (PI. n., fig. 15). The canal system is sy conoid, though not in a very typical way. The dermal cortex is very strongly developed. The dermal pores are circular and small, measuring about 60 /^ across on an average. The flagellate chambers, which are radially arranged around the gastral cavity, extend through the greater part of the thick- ness of wall. They are rather narrow, not quite straight, but more or less crooked ; most of them branch once or twice in the distal parts. The position of nucleus in collar cells is apical. The gastral cortex is much thinner than the dermal. It forms a continuous lacework, the angular meshes (150-300 jj- across) of which are visible to the naked eye. The meshes referred to con- stitute the openings of exhalant canals into the gastral cavity. The exhalant canals, arising from a single or more flagellate chambers, are short. Skeleton (PL H., fig. 15). The fairly thick dermal skeleton is made up of : 1) sub- regular or slightly sagittal triradiates, lying tangentially in several layers in a rather confuse arrangement ; 2) the paired rays of subdermal pseudosagittal triradiates ; and 3) small oxea grouped into small tufts which project from the sponge surface. ON SOME JAPANESE CALCAREOUS SPONGES. 25 The tubar skeleton is of the inarticnlate type, being composed of the strong basal rays of subdermal pseudosagittal triradiates as well as of subgastral sagittal triradiates of varying sizes, respect- ively directed centripetally and centrifugally. The gastral skeleton is thinner than the dermal. It is com- posed of the strong paired rays of subgastral sagittal triradiates, and of sagittal triradiates which are similar to those of the dermal skeleton except in being slightly smaller and rather more regularly arranged. The skeleton of the oscular margin is composed of oxea and triradiates, both placed densely together. The former run longi- tudinally and parallel with one another, the latter have strongly divergent paired rays. Spicules. Dermal triradiates (textfig. 5, a, h). — Subregular or slightly sagittal, varying in size. All rays conical, nearly straight, often slightly irregularly contoured, tapering from base to sharply pointed end, 130-410,« long and 20-50;« thick. Towards the osculum the triradiates become distinctly sagittaUy differentiated. Subdermal triradiates (textfig. 5, c, d). — ^Pseudosagittal, irregu- lar, strongly developed. All the rays differing in length, some- what irregular in contour. Basal ray almost straight, tapering from base to sharp point, occasionally slightly thicker than paired rays, 320-630 /^ long and 30-60 ,« thick. Paired rays are equal in thickness, but unequal in length and shape. The longer ray usually curved, sometimes crooked, gradually tapering, 160-430/^ long and 30-50 /^- thick. The shorter ray less curved than the longer, gradually pointed, 120-330 ß long and 30-50 « thick. Subgastral triradiates (textfig. 5, e,f). — Strongly sagittal. On 2é ART. 5. — SANJI HÔZAWA: Textfig. 5. Grantessa mitsukuni. a, b, Dermal triradiates. c, d. Subdermal triradiates. e, f, Subgastral triradiates. g, h, Gastral triradùites. i, j, k, Oxea. 1, m, Triradiates of oscular margin, n, o, p. Oxea of oscular mai^n. (AU figs. 100 X). the whole, they are about equally well developed as subdermal pseudosagittal triradiates. Basal ray almost straight, gradually sharp-pointed, slightly thicker than paired rays, 240-580 fJ- long and 20-50 /^ thick. Paired rays almost equally long, tapering from base to sharp point, 130-310/^ long and 16-50;" thick. Gastral triradiates (textfig. 5, g, h). — Slender. Basal ray quite straight, tapering in their entire length, sharply pointed, 120-220 /^ long and 16-24 ;" thick. Paired rays curved at base, gradually ON SOME JAPANESE CALCAREOUS SPONGES. 27 sharp-pointed, nearly as long as basal ray, 160-200 /-« long and 16-24;« thick. Oxea (textfig. 5, /,/, A). — Very small, slender, straight, broadest nearer one end than the other, tapering towards both pointed ends. One example measured was 130 ;« long and 4 /^ in the thickest part. Triradiates of oscular margin (textfig. 5, /, m). — Basal ray quite straight, gradually tapering, sharp-pointed, 130-250 /^ long and 12-16 fi thick. Paired rays stouter and shorter than basal ray, nearly at right angles to the latter ; straight or slightly curved, 120-220 i« long and 16-20// thick. Oxea of oscular margin. — There exist two kinds of oxea. The thicker kind is irregularly curved and rather bluntly pointed at ends ; sometimes provided with a nodiform ring at the free end, 250-470^ long and 16-20;" thick (textfig. 5, ?z, o, ;:»). The thinner kind resembles oxea of dermal cortex, but is longer ; broadest near the inner end, 300-460 /^ or more long and 4-6 /^ thick. Note. — This interesting species seems to be closely related to Grantessa sycilloides (Schuffnee)^^ of the Indian Ocean, but can be distinguished from it cliiefly by the external appearance, by the presence of thin dermal oxea, by the flagellate chambers being laterally not fused with one another, and by the basal rays of subgastral sagittal trii^adiates being equally developed as in those of subdermal. The species is named after the late Professor MiTSUKUEi, the collector of the type -specimen. Locality. — Koajiro Misaki. 1) Sycortls sycilloides Schuffxee, 1887 (1), p. 420, Taf. XXV, Fig. 10. 28 ART. 5. — SANH HÖZAWA : Genus Heteropia Cakter (omend.). 6. Meteropia striata, u, sp. (PL I., fig. 8 ; PI. n., fig. 16 ; textfig. 6). This new species is based on three specimens in the Science College. Two of them were obtained by the late Professor MiTSUKURi at Koajiro, near the Misaki Marine Biological Station. A third specimen was collected by myself at Aburatsbo, also close to the Misaki Marine Biological Station. It was found attached to the under side of a floating log. The first specimen (Sei. Coll. Spec. No. 25 ; PL I., fig. 8), which is herewith made the type of the species, is a small colony of about eighteen tubular individuals, united together at their base and most of which are bent in the same direction. They are broadest at base and taper distally. The larger individuals are provided with terminal osculum fringed with very feebly developed trichoxea ; the smaller individuals show neither osculum nor the fringe. The largest individual measures 8 mm. in total length and 3 mm. in greatest breadth, the wall reaching about 0.7 mm. in thickness. The osculum is circular with a diameter of about 0.6 mm. The surface shows longitudinal striation due to the presence of large oxea in dermal cortex, and is hispid on account of vertically projecting hair-like oxea. Colour in alcohol is white ; texture rigid. The second specimen (Sei. Coll. Spec. No. 20) closely resembles the first in both external appeai'ance and microscopical structure. The third specimen (Sei. Coll. Spec. No. 27) is the largest of all. The colony is composed of numerous branching and anasto- mosing tubes of varying cahbre. Most of the tubes are provid- ON SOME JAPANESE CALCAEEOUS SPONGES. 29 cd at their free end with a circular oscnlam of J-1 J mm. dia- meter ; the marginal fi-ingo is scarcely visible to the naked eye. The tubes attain a breadth of about 5 mm. ; thickness of wall about 1 mm. The colour of the sponge in the preserved state is brownish white. The following description is based on the type- specimen. Canal system (PL IL, fig. IG). The canal system of this species is not of tlie true syconoid type, unlike others of the genus. It is rather of an intermediate type between the sylleibid and the leuconoid. The flagellate chambers vary much in shape and size, fi'om those of spherical shape measuring 50-100 i>- diameter to others of an elongate sac-like configuration, say, 200 /^ by 80 /^ in dimension. They are closely set in the chamber layer, showing a somewhat radial— though not quite strictly radial — arrangement around wide and long, sometimes slightly branched exhalant canals. The posi- tion of nucleus in the collar cells is apical. The exhalant canals perforate the thin gastral cortex and open into the gastral cavity by angular mesh-like pores of various sizes, up to 150-250/^ in length. Skeleton (PI. H., fig. 16). The fairly thick dermal skeleton consists of : 1) small sagittal triradiates, lying tangentially in several layers, with their basal rays generally directed towards sponge base ; 2) large oxea run- ning longitudinally ; and 3) the paired rays of subdermal pseudo- sagittal triradiates. Besides these, there may occur slender hair- like oxea, very sparsely distributed and projecting on the dermal surface. 30 AKT. 5. — SANJI HÔZAWA : The tiibar skeleton may be said to be on the whole of the inarticulate type. It is chiefly composed of the basal rays of subdermal pseudosagittal and subgastral sagittal triradiates. The oppositely directed rays in question lie side by side almost in their entire length. Usually the subdermal triradiates are less developed than the subgastral. The gastral skeleton is much thinner than the dermal. It is made up of the paired rays of subgastral sagittal triradiates, of sagittal triradiates placed tangentially and with their basal rays pointing towards sponge base, and of a few quadrh'adiates which have their short apical ray projecting into the gastral cavity. The skeleton of oscular margin is, in full-grown individuals, composed of oxea and of triradiates, both closely set. The former are of two kinds : the one is similar to that found in dermal cortex, while the other is very thin and hair-like. The oscular triradiates, which are apparently differentiated from the gastral triradiates, have very slender basal ray and the stouter paired rays standing nearly at right angles to it. Spicules. Dermal triradiates (textfig. 6, a, h). — Slightly sagittal. All rays of equal thickness, tapering fi-om base to sharp point. Basal ray quite straight, 80-170 /^ long and 8-12 p. thick. Paired rays of about the same length, almost straight, 60-100 ;" long and 8-12 /^ thick. Subdermal triradiates (textfig. 6, c, d). — Pseudosagittal, stout, all rays of different length but of nearly the same tliickness, not quite smooth. Basal ray generally curved near base, gradually and sharply pointed. Paired rays of different length and shape. The longer ray is slightly shorter than the basal ray, almost ON SOME JAPANESE CALCAREOUS SPONGES. 31 Textfig. 6. Heteropia striata. a, b, Dermal triradiates. c, d. Subdermal triradiates. e, f, Subgastral triri^diates. g, h, Gastral triradiates. i, j, Gastral qnadriradiates. k, 1, m. Dermal oxea. (a— j, 150 X ; k— m, 40 x ). straight excepting the sHght curvature near base, gradually taper- ing to a sharp point. The shorter ray almost straight, tapering from base to sharp point. In a typical case, the basal ray is 190 /^ long ; the longer paired ray 130 ;" long ; and the shorter paired ray 110/^ long; all 16 fi thick. Subgastral triradiates (textfig. 6, e,/). — Sagittal, with sharply pointed rays of nearly equal thickness. Basal ray straight, taper- ing from base to sharp point, 150-300^ long and 16-20/^ thick. 32 ART. 5. — SANJI HÔZAWA : Paired rays much diverging, almost of equal length, slightly curved, 90-170 // long and W-20 ft thick. Gastral triradiates (textfig. 6, g, h). — Sagittal. Basal ray usually much longer and slightly thinner than paired rays, quite straight, tapering gradually from base to sharp point, 170-230 /^ long and 8-12 // thick. Paired rays almost of equal length, straight or shghtly curved, gradually and sharply pointed, 90-150 ,« long and 12-16/^ thick. The spicules closely resembL? dermal triradiates but may be distinguished from these by the larger size, by the basal ray being thinner than paired rays, and by their more regular arrangement. Gastral quadriradiates (textfig. C, i,J). — Similar to gastral tri- radiates, except in the presence of apical ray. Apical ray very short, thinner than either of the basal or i)aired rays, slightly curved and gradually sharp-pointed, 40-60 /^ long and 6-8 /^ thick. Large dermal oxea (textfig. 6, k, I, m). — Strongly developed, spindle-shaped, a little irregular in outline, generally broadest at a point nearer proximal than distal end and tapering towards both sharply pointed ends. They are more or less curved, and are of very variable length, 0.49-2 mm. long and 30-90 [j- thick. Trichoxea of dermal cortex. — Hair-like, straight or shghtly curved, generally with the free end broken off, 2-4 « thick. A large example measured 700 /^ in length and 2 /^ in tliickness. Trichoxea of oscular margin. — Nearly like those of dermal cortex. Hei'e again, the free end is usually found broken off. An example, incomplete at one end, measured 240 /^ in length and 4 // in thickness. Note. — I have referred this new species to the genus Heteropia, chiefly because of the presence in dermal cortex of the large oxea running nearly parallel with the surface. Remarkable is the non ON SOME JAPANESE CALCAEEOUS SPONGES. 33 syconoid type of its canal system, in which respect it differs from all other members of the genus. It is evidently a species closely related to Heferopia glomerosa}^ Localities. — Koajiro and Aburatsbo, near the Misaki Marino Biological Station. Genus Amphiute Hanitsch. 7. AmpJiiute ijiinai, u. sp. (PI. I., fig. 9 ; PL n., fig. 17 ; testfig. 7). This new species is represented in the collection by three specimens, aU obtained at Döketsba, Sagami Sea, from a depth of 215-257 m. The first specimen (Sei. Coll. Spec. No. 18), which I make the type of the species, is a single person of a somewhat curved and laterally compressed, elongate cylindrical form, broadest at a part a little above the middle. The total length is about 60 mm. and the greatest breadth 12 mm. The osculum is in part damaged, besides being in a collapsed state. It shows neither an oscular fringe nor a collar. The gastral cavity is deep and extends throughout the entire length of the sponge. The lower part of the sponge is drawn out in a stalk-like manner, the body in this region being narrowed to a breadth of 5 mm. The dermal surface of the sponge, except in the lower parts, is smooth and shows fine longitudinal striation due to the presence of large oxea in dermal cortex. Thickness of the wall, as meas- ured in the broadest part of the sponge, is about 1 mm. The 1) Leuronia (jlomerosa Boweebank, Proc. Zool. Soc. London, 1873, pp. 17-19, PI. IV, Figs. 1-G; IMeroim. glomerom, Dendy, 1î)15 (4), pp. 83-86, PI. I, Figs. 3, 3a, 3b; PI. H, Figs. 8a-8g. 34 AKT. 5. — SANn HÔZAWA : colour in alcohol is greyish white. The texture is rather soft and not very compact. The second specimen (Sei. Coll. Spec. No. 19) consists of a single cylindrical person with the lower parts torn off. It is 37 nmi. long and 10 mm. broad in the broadest part, where the thickness of waU is less than 2 mm. The osculum, turned towards one side, is irregular in shape and measures about 3 mm. across in a way. Its margin is very thin and without an obvious fringe or collar. The colour in alcohol is greyish white ; tlie texture is firm and elastic. The third specimen (Sei. Coll. Spec. No. 28 ; PI. I., fig. 9) is rather spindle-shaped and shghtly laterally compressed. It meas- ures 80 mm. in total length, 12 mm. in greatest breadth, and about 1 J mm. in maximum thickness of wall. The osculum at the upper truncate end is almost naked, with thin irregularly undulating margin. It leads into a wide and deep gastral cavity. The narrowed stalk-like base of the sponge is provided with a number of small irregular processes for attachment. The colour in alcohol is white and the texture is fairly firm and elastic. The following description is based on the type (first) speci- men. Canal system (PI. n., fig. 17). The canal system stands somewhat intermediate between syUeibid and leuconoid types, approaching more nearly the latter than the former. The dermal pores, measuring 80-150 /-« in dia- meter, are closely distributed all over the surface. Canals starting from several pores join together to form very wide inhalant canals which run deep into the wall. The exhalant canals are also very wide and extend through the greater part of the wall thickness. ON SOME JAPANESE CALCAREOUS SPONGES. 35 They open into the gastral cavity by apertures of 150-250 /^ dia- meter, separated by interspaces of 50-250 /^. The flagellate chambers, densely arranged between inhalant and exhalant canals, are generally of an oval shape, about 80- 150// in the shorter diameter. Some of them, especially those in the periphery of the sponge, may be of a more or less elongate shape and are arranged radially around the exhalant canal, thus suggesting the sylleibid condition. Skeleton (PI. II., fig. 17). The dermal skeleton is composed of microxea, triradiates, large oxea and the paired rays of subdermal pseudosagittal tri- radiates. Nearly all tlie microxea stand vertically to the dermal surface. Those at the oscular margin stand out parallel with the long axis of the sponge. The trii'adiates are placed tangentially in a few layers, with the basal ray pointing more or less down- wards. The large oxea generally run longitudinally, covering all over the sponge surface. The tubar skeleton is of the inarticulate type, being composed of the centripetal rays of subdermal pseudosagittal triradiates and of the centrifugal rays of subgastral sagittal triradiates. The ends of the rays mentioned usually reach beyond the middle of the w^all, and there exist between them no intermediate spicules in- dicative of an articulate skeleton. The gastral skeleton is thinner than the dermal and contains tlie paired rays of subgastral triradiates as well as large oxea and quadriradiates. The oxea nearly resemble those of the dermal cortex, but are much less numerous and somewhat more irregularly placed. The quadriradiates are slender-rayed and fairly large in 36 ART. 5.- — SANn HÔZAWA : number. Their basal rays point downwards, while the apical rays project into the gastral cavity. Spicules. Dermal triradiates (textfig. 7, a, h). — Typically slightly sagittal. Basal ray straight, gradually tapering and sharply pointed, either slightly longer or shorter and a little more slender than the paired rays, 80-190 [j- long and 8-12 /^ thick. The paired rays subequal in length, generally straight excepting a slight curvature near base, sometimes a little crooked, gradually tapering, sharp-pointed, 80- 180 A« long and 12-14// thick. Subdermal triradiates (textfig. 1 , c, d, e). — Pseudosagittal. All rays of different length and shape. Basal ray longer than the paired rays, straight except being slightly bent near base, gradually sharp-pointed, 280-600,« long and 14-16 /^ thick. The shorter of the paired rays is nearly straight, gradually tapering and sharply pointed, 80-170 // long and 12 // thick. The longer of the paired rays shghtly bent near base, sometimes crooked farther out, gradu- ally and sharply pointed ; they are as thick as the basal ray and in most cases thicker than the shorter of the paired rays, 150-280/-/ long and 14-16/^ thick. Subgastral triradiates (textfig. 7,/, (j). — Strongly sagittal, with the paired rays very strongly divergent. Basal ray quite straight, and gradually sharp-pointed, 380-880 // long and 12 // thick. Paired rays almost as thick as the basal ray, equal or shghtly differentiated in length, gently curved or crooked, gradually tapering and sharply pointed, 170-220// long and 12/^ thick. Gastral quadriradiates (textfig. 7, h, i). — Slender ; facial rays sagittal in most cases. Basal ray generally the longest, quite straight, distinctly tapering in the basal parts but less so in the ON SO:\IE JAPANESl-^- CALCAREOUS SPONGES. 37 Textfig. 7. Amplmde ijimai. a, b. Dermal triradiates. c, d, e, Subdermal triradiates. f, g, Siibgastral triradiates. li, i, Gastral quadriradiates. j, k, Large dermal oxea. 1, m. Large gastral oxea. o, p, q, Microxea. (a — i, o — q 100 X ; j — m, 40 x ). remaining greater parts, 230-360 /^ long and 8 /i thick. Paired rays slightly curved, a little thicker than the basal, with a wide oral angle, generally equally long but sometimes differentiated 38 AET. 5. — SANJI HÔZAWA : in length, nearly uniformly thick in the greater parts of their length, pointed at end, 90-230 /^ long 12// thick. Apical ray almost straight, sometimes sHghtly crooked in the apical parts, gradually sharp -pointed, under 150// length and 8/^ thickness. Large oxea of dermal cortex (textfig. 7,/, k). — Usually large, straight or shghtly curved, thickest at about the middle, gradually tapering towards sharp-pointed ends, 1.2-3 mm. long and 50-80 /^. thick. Large oxea of gastral cortex (textfig. 7, /, r>i).— Nearly similar to those of dermal cortex, being differentiated from these only in the fewer number and in the less reo;ular arrangement. Microxea (textfig. 7, o, j?, q). — Generally slightly curved, nearly uniformly thick throughout their length, and pointed at both ends, 110-210// long and 6-8// thick. Note. — This sj^ecies appears to be referable to AmpMufe Hanitsch, by the presence of subdermal pseudosagittal and sub- gastral triradiates and of the large longitudinal oxea in both dermal and gastral cortices. But it differs in several respects from the only other known species of the genus Amphiute paulini, which was first described by Hanitsch (1, 2) and afterwards also by Beeitfuss (4) from the west coast of Portugal. The difference consists chiefly in the present form being sohtary, in the osculum being destitute of a distinct and conspicuous fringe, and in the dimensions and other details of the characters of most spicules. I take pleasure in naming this interesting species after Prof- essor I.JIMA. Locality. — Doketsba, Sagami Sea. ON SO:SŒ JAPANESE CALCAREOUS SPONGES. 39 List of References. Arnesen, E. (1) (1901). Spongier fra deu norske kyst. I. Calcarea. Systemmatlsk katalog med bemerkinger og bestemmelsestabel. Bergens Mus. Aarbog, 1900 (1901), No. 5. Bowerbank, J. S. (1) (1872-1876). Contributions to a General History of the Spongiadœ. Proc. Zool. Soc. Loudon, 1872, pp. 115-129, 196-202, 626-634, PI. V, VI, X, XI, XLVI-XLIX ; 1873, pp. 3-25, 319-333, PI. I-TV, XXVni-XXXI; 1874, pp. 298-305, PI. XLVI, XLVH; 1875, pp. 281-296; 1876, pp. 768-775, PI. LXXVni-LXXXI. Breitfuss, L. (1) (1896). Kalkschwämme der Bremer-Expedition nach Ost-Spitzbergen im Jahre 1889 (Prof. W. Kükenthal und Dr. A. Walter). Zool. Anzeiger, Bd. XIX, No. 514, pp. 426-432. . (2) (1897). Catalog der Calcarea der zoologischen Sammlung des königlichen Museums für Natm'kunde zu Berlin. Arch, füi' Naturgesch., Jalu'gang LXm, Bd. 1, pp. 205-226. . (3) (1898). Kalkschwammfauna des weissen Meeres und Eismeerküsten des em-opäischen Russlands mit Berücksichtigung und Aufstellung der Kalkschwammfauna der arktischen Kegion. Mémoires de l'Acad. Im- periales Sciences, St. Petersbourg, (Ser. 8), Vol. VI, No. 2, Taf. I-IV. . (4) (1898). Kalkschwammfamia der Westküste Portugals. Zool. Jahrb. Syst. Abt. Bd. XI, pp. 89-102, Taf. XI. . (5) (1898). Kalkschwammfauna von Spitzbergen. Nach dem Samm- hmg der Bremer-Expedition nach Ost-Spitzbergen im Jahre 1889 (Prof. AV. KÜKENTHAL und Dr. A. Walter). Zool. Jahrb. Syst. Abt. Bd. XI, pp. 103-120, Taf. Xn, XIH. Carter, H. J. (1) (1885-1886). Descriptions of Sponges from the Neigh- bom-hood of Port PhiUip Heads, South Australia. Ann. Mag. Nat. Hist., (Ser. 5), Vol. XVH, pp. 431-441, 502-516; Vol. XVEŒ, pp. 34-55, 126-149. 40 ART. 5. — SANJI HÔZAWA : Carter,' H. J. (2) (1886). Description of n new species {Aphroceras ramosa). (in Higgin, T., Report on the Porifera of the L.M.B.C. dis- trict). Proc. Lit. PhiL Soc. Liverpool, Voh XL. Appendix. Dendy, A. (1) (1892). Synox^sis of the AiistraHan Calcarea Heterocœla ; with a proposed Classification of the Group and Descriptions of some New General and Species. Proc. Roy. Soc. Victoria, (n.s.), Vol. V, pp. 69-116. . (2) (1893). Studies on the Comparative Anatomy of Sponges. V. Observations on the Structm'e and Classification of the Calcarea Hetero- cœla. Quart. Joura. Microsc. Sei., (n.s.), Vol. XXXV, pp. 159-257, PI. X-XIV. . (3) (1913). Report on the Calcareous Sponges collected by H.M.S. " Sealark " in the Indian Ocean. Trans. Linnean Soc. London, Zool. Vol. XVI. Part. I, pp. 1-29, PI. I-V. . (4) (1915). Report on the Calcareous Sponges collected by Mr. James HoRNELL at Okhamandal in Kattiawar in 1905-6. Report to the Government of Baroda on the Marine Zoology of Okhamandal in Kat- tiawar, Part n, 1915. Dendy, A. and Row, W. H. (1) (1913). The Classification and phylogeny of the Calcareous Sponges, with a Reference List of all the described Species, systematically aiTanged. Proc. Zool. Soc. London, 1913, pp. 704-813. Haeckel, E. (1) (1872). Die Kalkschwämme, eine Monographie. Berlin, 1872. Hanîtsch, R. (l) (1894). Amphiute, eine neue Gattung lieterocœler Kalk- schwämme. Zool. Anzeiger, Vol. XVII, p. 433. . (2) (1895). Notes on a Collection of Sponges from the west coast of Poi-tugal. Trans. Liverpool Biol. Soc, Vol. IX, pp. 205-219, PI. XH, xin. Jenkin, C. F. (1) (1908). The Marine Fauna of Zanzibar and British East Africa., from Collections made by Cyril Crossland, M. A., in the Years 1901 and 1902. The Calcareous Sponges. Proc. Zool. Soc. London, 1908, pp. 434-456. ON SOME JAPANESE CALCAREOUS SrONGES. 41 Lambe, L. M. (1) (1900). »Spouges from tho ( '(Uist of North-eastern Oiuada, and Greenlaud. Trans. JÎ03'. Soc. Canada, (Ser. 2), Vol. VI, S(«ct. 4, i)]). 19-49, PI. I- VI. Lendenfeld, R. von. (1) (1885). A Monograph of the Austi-ahan Sponges. Part HI. The Calcispongiee. Proc. Linn. Soc. New South Wales, Vol. IX, pp. 1083-1150, PI. LIX-LXVn. Polejaeflf, N. (1) (1883). The Calcarea. Report on the Scientific Ptesnlts of tlie A'(jyage of H.M.S. Challenger, Zoology, Vol. WH. Preiwisch, J, (1) (1904). Kalkschwämme aus dem Pacific, li-gebnisse «iner Reise nach dem Pacific, Schauinsland 1896-1897. Zool. Jaln-b. Syst. Alit., Bd. XIX, pp. 9-26, Taf. n-IV. Bidley, S. 0. (1) (1881). Spongida, Account of the Zoological Collections made during the Survey of H.M.S. Alert in the Straits of Magellan and on the Coast of Patagcmia. Proc. Zool. Soc. London, 1881, pp. 107-137, PI. X, XI. Row, W. H. (1) (1909). Report on the Sponges CoUected by Mr. Cybil Crossland in 1904-5. Part I. Calcarea. Jom-u. Linn. Soc. London, Zoology. Vol. XXXI, pp. 182-214, PI. XIX, XX. Schuflfner, 0. (1) (1877). Beschreibung einiger neuer Kalkschwämme. Jenaische Zeitschr. Vol. XI, pp. 403-433, PL XXIV-XXVI. Published September SOth, 1916. s. IIÖZAWA: ON SOME JAPANESE CALCAREOUS SPONGES. PLATK I. Plate I. Grantcssa sldmeji, u, sp. ; natural size. Gnmtœsa sldmeji, n. sp. ; natural size. Gntntessa sagamiana, ii. sp. ; natural size. Grantcssa intusarticulata (Carter) ; natural size. Grantessa. intusarticulata (Carter) ; natural size. Grantcssa hasipapillata, n. sp. ; natural size. Grantessa mitsuhurii, n. sp. ; natural size. Fig. 8. Hcteropia striata, n. sp. ; natural size. Fig. 9. Amphiutc ijimai, n. sp. ; natural size. Fi^^ 1. Fig. 2. Vi^. 8. Fig. 4. Fig. 5. Fig. 0. Fig. 7. Jour. Sei. Coll., Vol. XXXVlll.. Art. 5. PL. I. S. llomu'a : Jopanei^e Calcareous; Sponge.i, - 1 s. irOZAWA: SOME JAPANESE CALCAREOUS SPONGES. PLATE II Plate II. Fig, 10. Grantessa skimeß, n. sp. Part of a horizontal section ; 150 x Fig. 11. Grantessa s/dmeji, n. sp. Pai-t of a horizontal section ; 150 x Fig. 12. Grantessa sagamiana, n. sp. Pai-t of a horizontal section ; 100 x Fig. 13. Grantessa intusarticulata (Caeter). Part of a horizontal section ; 100 x Fig. 14. Grantessa hasiiMpillata, n. sp. Part of a horizontal section ; 100 x Fig. 15. Grantessa mitsuhurii, n. sp. Part of a horizontal section ; 100 x Fig. 16. Heteropia striata, n. sp. Part of a horizontal section ; 150 x Fig. 17. AmjjJttute ijimai, n. sp. Part of a horizontal section ; 100 x Jour. Sei. Coll., Vol. XXXVIII., Art. 5. PL II. S. HOiciwa : Japanese Calcareous Spotiges. JOURNAL OF THE COLLEGE OF SCIENCE, IMPERIAL UNIVERSITY, TOKYO. VOL. XXXVIII., ARTICLE 6. Injection Experiments on Plants. By Yasutaro YENDO, RujakusM, NögakmM. With tioo Plates. I. Introduction. The first experiments of injections into the plant body were made by Ekhaet (1) and Reichakd. They experimented with 8 species of plants, employing various solutions, and showed the possibility of introducing by injection extraneous substances into the sap of the plant body although the efïects varied according to the material employed. Some experiments have recently been made for particular purposes, in which the injection method was used. Weber (2) succeeded, for example, in influencing the unfolding of winter buds by means of a water- injection, Jesenko (3) made similar researches, employing certain chemical substances. On the whole, the investigations which hitherto have been made on this subject are scanty, but interesting in the results. As Peof. Miyoshi (4) has remarked, the organisation of plants differs from that of animals in respect to the imperfeetness of the circulating system and the higher grade of individuality of several parts, so that we cannot expect to apply the injection 2 ART. 6. — Y. YENDO : method to a plant so effectively as is done with animals. If we should succeed, however, in making particular chemical substances circulate in a certain measure through the body of plants, it might be possible to stimulate their development or to cure them of diseases, and moreover, to render them immune by the injection of certain substances. The scope of the present study is to answer the following questions : — 1) Is it possible that a certain amount of a substance injected into a certain part of the plant body, is conducted through the entire plant ? In other words, how far and to what part is an injected solution conducted during a definite time ? 2) How does the effect of injection differ according to the kinds of plants, organs and tissues ? The present investigation was carried out at the suggestion and under the guidance of Peof. M. Miyoshi, during the academic year of 1911-12 in the Botanical Institute of the College of Science, Imperial University of Tokyo. I beg here to express my gratitude to him for his generous assistance and advice. II. Method of injection. A medical syringe, provided with measured shaft and 1.5 c.cm. in volume, was employed. Usually the injection was made by inserting the cannula into the plant body deej^ly and obliquely, until it touched a vascular bundle ; in the case of injecting into a hard tissue of a plant, a slit was made with a knife and then the syringe was applied. We have to ascertain into what parts the injected solutioii penetrates and how far the fluid reaches. On injecting a coloured solution I observed that it enters the air- containing intercellular INJECTION EXl'KimiENrs OX Pr.\NTS. o spaces and the lumen of vascular bundles more easily than tlie interior of the cells. The range attainable by the injected solution may differ in accordance with the structure, size and nature of the plant, the direction of the injecting needle applied, the intensity of the pres- sure of injection, the amount of the solution injected, etc. I found by my experiments that the injected solution, 0.03 c.cm. — 0.15 c.cm. in volume, does not extend far from the point of injection, but is arrested within a few centimeters. III. Experiments with various injection solutions. The following aqueous solutions were used in my experi- ments : — lOX. 1) Lithium nitrate. 1%', 5^ 2) Copper sulphate. 1% ; 5X 3) Eosin, 0.5^. 4) Aniline violet. 0.2X. Lithium is a metal which is not usually found in the plant body, but there are some plaats containing it normally (Tscher- MAK 5), hence in experimenting with a lithium salt solution, we must take care to choose a plant which is free from it. Lithium salts act poisonously on several plants (Gtauneesdoefer, 6) even in small quantities, but as its toxicity is not so intense as that of SO^Cu or eosin, and moreover, as it is quickly carried through the plant body, it may most suitably be used for our purpose. A minute amount of lithium is readily detectable spec- troscopically and we can roughly determine its amount from the intensity and duration of the red lithion line (Teuchot, 7). Obser- vation has shown that it is sometimes obscured by the presence of too many other metal-lines appearing in the spectrum. AUT. 6. YEXDO : Though I mainly used solutions of lithium nitrate, those of SO^Cu, eosin and aniline violet were also employed for the purpose of tracing microscopically the path through which they are con- ducted, and to compare the modes of conduction. IV. Injection experiments on several plant-organs. Oi-gan injected Plant name No. of exp. Remarks Root Mhaphanus sativus 67,71 conical, fleshy. J' Taraxacum officinale 10 contai ILS milky jiiico. Rhizome Carex pumila 25 jointed. }) Menianthes trifoliata 28 lacunose. )} Calystegia soldanclla 13 contains milky juice. }y Equisetiim arvense 42 >} Nuphar japonicum 32 submerged. Bulb Karcissus pjapyracevs 22 coated. }f Solanum tuberosum 11 a tuber. Stem Cornus contvoversa 5 woody. >} Gamut lutea 18 Setaria sp. Hicinus communis 19 6 a culm. jointed, pro\àded with medullary hole. a }} Cosmos bipinnatus 5 }> IViehosanthes cucmneroides 15 succulent, climbing. )} Pucvaria Th unbergiana 17 liard, climbing. >.» Lycojwdium. clavaturn 41 creeping. 77 Potamogcton poli/gonifolius 37 slender, submerged. 77 MyriophyUum spicatum 34,35 do. 77 Cereus serpentinus 61 fleshy. Cotyledon T7c/« Paba 12 fleshy. Petiole Magnolia hypoleuca 4 77 Aspidium lacerum 39 INJECTION KNPEnnrENTS ON Pr.ANTS. Potiolc Traclii/carjms execha 60 Loaf Allium fistalosum 62, 69, 72 ,, Rhoco discolor 20,21 >f Narcissus pa2-}yraceui^ 23 » Cotyledon inalacojihyUa 9 succulent. Iiitlo- rescencc Typlia japonicci Fatsia japonica 36 1 lacunose. )) Narcissus papyraccus 24 » Nuphar japonicnm 30 Fruit Citrus nobilis 3 a iiespkliuiu. V. Injection into parenchymatous tissue. (Compare exp. 77-83.) For these experiments, we must choose plants with thick parencliymatous tissue, so that the injection may be performed witli ease. Wlien injecting into tlae cortical parenchymatous tissue, the needle was inserted as little as possible, nearly parallel to the surface of that part, and then the solution was gently injected. Injection into the medullary parenchymatous tissue often proves difficult as it is necessary to expose a part of the pith by cutting out some parts of the outer tissues (cortical tissues and vascular bundles) with a sharp knife. The needle must be introduced deeply into the pith and brought as nearly parallel as possible to the axis of this ; then the injection should take place gently and with great care, so as to avoid over- injection into other parts. For a plant having a wide medullary cavity (e.g. Vicia Faba), there may be no need of cutting off the outer tissues, since the medullary air-space readily takes up the solution injected. In this case, insert the tip of the cannula into the medullary cavity, piercing the body G ART. 6. — Y. YEXDO : wall of the plant, and then inject the solution very slowly so as to avoid excess. I experimented with Vicia Faha (exp. 79), applying the aniline violet solution, and observed that the solution injected into the medullary cavity was at first absorbed by the innermost layer of pith-cells and was then carried by degrees through the pith-cells toward the vascular bundles, so that it penetrated almost all the pith and xylem parts. The walls of the xylem elements were especially well tinged by it (see PI. II, Fig. 2), while the outer tissues (i.e. ' cambium, phloem and primary cortex) remained entirely free of it. This proves that the dye is carried to a distance only through the xylem elements. In the case of lithium solutions applied, I invariably observed that lithium was conducted through the cortical and medullary parenchymatous tissues, but it is certain that the longitudinal conduction is accomplished by the vascular bundles. We may affirm from these facts that the parenchymatous cell is capable of conducting certain injected substances, but its power of conduction is feeble in comparison with that of conducting tissues proper. VI. Experiments in the moist chamber. I made some injection experiments in the moist chamber with Vicia Faha. The results (see exp. 48a, 49) agreed in all cases ; namely, the injected substance was carried not only upwards but also downwards from the stem into which it was intro- duced, and penetrated the branches attached to the stem at tolerable distances below the point of injection. Lithium was detect- ed, in this case, in the petioles and midribs of most of the leaves, but never in the blades (small veins and mesophyU). In the control plant (exp. 48b), however, it was present especially in the INJECTION EXPERIMENTS ON I>r>ANTS. 7 blade, with the exception of young terminal leaves, and the downward conduction did not occur conspicuously as a rule. This marked difference must be due to the absence of transpiration. VII. Experiments on some deciduous trees in winter and early spring. (See exp. 50-58.) In order to observe the mode of conduction in deciduous trees in winter and early spring, some injection experiments were per- formed. From these experiments it was found that lithium w^as carried upwards to a considerable distance from the injected point, but downwards only to a comparatively short distance ; the con- duction in these plants, however, is feebler than that in leafy plants. I observed in Cornus confroversa that the conduction of lithium gradually increased in early spring ; in the leafless plant in winter (exp. 53-55) hthium was not carried far, even after a considerable lapse of time ; when the plant was at the bleeding period in early spring and the buds had unfolded a little, an increase in the rate of conduction was observed (exp. 56), but lithium did not yet reach the buds (PI. II, Fig. 5. A). Only when the young leaves attained about 3 cm. in length, (PI. II, Fig. 5. B) it was found in them for the first time (exp. 57). We may infer from these ex- periments that transpiration plays the most important rôle in the conduction of the injected substance, although osmotic action of cells and the root-pressure etc. have more or less influence. VIII. Experiments on some Thallophytes. As will be seen in the experimental data (43, 44), certain fungi distinctly showed the conduction of lithium through their 8 AKT. 6. Y. YEXDO : body. In Cortinellus edodes (exp. 45) I observed that the upward conduction was more extensive to the downward one and that the transverse conduction occurred within a certain Hmit. In Usnea barbata v. scahrosa (exp. 46), no conduction of hthium was observed. As regards algae (Oltmanns, 15), most of them are too compactly structured and their cells filled with sHmy or watery contents, so that there exists in them no space to receive injections, making them unsuited for injection experiments. Eisenia and EcUonia were employed in my experiments, since they have rather loosely con- structed medulla. The algae were examined immediately after the injection, and I have found that hyphae, siebhyphae as well as the slimy substance, were well tinged in the injected part of the medulla ; and in the ease of eosin injection the contents in the siebhyphae were espe- cially deeply coloured. The algae which were subjected to injection were kept in a sea-water reservoir at the sea-shore during the experiment. The results indicated that the injected pigment gradually dis- coloured in the course of time, and after a few days it was found very slightly in the injected parts, the hyphae as well as the shmy substances were almost entirely gone, but frequently a few inner rows of the cortical cells were found slightly tinged. In using SO.Cu (exp. G3) and LiXOj as injection material, similar results were obtained. Bryopsis pulmosa, a unicellular plant, may be worth investiga- tion, to observe the conduction of injected substances. The injected colouring matter was visible through its transparent cell-wall ; the arrangement of the contents was disturbed and well permeated by the injected solution. The injected plants were kept under sea- water, and observations were made at intervals. JXJKCTJON KXI'KUnriONT.S OX PLAXTH. 9 The result showed that conduction of the injected substances could not be seen, and they discoloured in the course of time (exp. 64, 70, 73). The lithium injection was also performed with this plant, but no conspicuous transportation was observable (exp. 64). In short, we conclude from these experiments, that no con- duction of the injected substance occurs, but merely diffusion, and some parts of the injected solution gradually pass out of the injected wound into the sea-water. IX. Discussion of the results. 1. Relation bet\\'eex traxspiratiox and coxductiox or INJECTED solution. The experimental data given below show that an injection solution, especially lithium, when injected into a stem is always carried toward the transpiring organs, the leaves ; but not so much to other organs, e.g. inflorescences (exp. 6, Ricinus communis, PI. I, Fig. 5), buds and young organs, e.g. young leaves, terminal buds, young shoots ; moreover, lithium when injected into a cotyledon, rhizome, bulb and root is carried toward the shoot. In the case of transpiration having been retarded, the conduction of the injected solutions always diminishes. For instance, Vicia Faha (exp. 48, 49) cultured in the moist chamber, as above mentioned, showed certain differences from the control plants. The experiments on deciduous trees in winter indicated that the conduction of lithium is feebler than that in leafy plants. It is manifest from these facts that transpiration plays an important rôle in conducting the injected solution. In the experiment with tlie inflorescence of Fafsia japonica 10 AßT. 6, — Y. yp:xdo : (exp. 1. PL I, Fig. 1), I distinctly found lithium in flowers which were just opening, but never in flower-buds and fruits. In Narcis- sus papyraceus (exp. 24), lithium was found in the uppermost opened flower, but scarcely in lower flower-buds which were covered with a broad bract. In these instances, we see that the injected substance is apt to be carried to the most transpiring part. I will now describe the results of the experiment on water- plants. When the terminal portion of Myriophyllum (exp. 35, PI. I, Fig. 7) emerged from the water, lithium when injected in the submerged part of the stem, was detected abundantly in the aerial parts, more especially in the leaves, whilst it was almost absent in the submerged ones. A little lithium was also found in the emerged inflorescence of Potamogeton crispus, but it was not detectable in the submerged leaves (exp. 38). Submerged plants, having no transpiring parts, also indicated conduction (Thoday, Skykes, Snell, 12, 13) of injected lithium ; in these plants, however, lithium was usually not detectable in the lamina even in the case of its evident presence in the petiole and midrib, for example, in submerged forms of Nuphar japonicum (exp. 29, 32). In the case of slender submerged forms, as Myriophyllum spicatum, I could not detect lithium throughout the body (exp. 34), and after a few days lithium was not found even in the injected part. I wished to ascertain whether these submerged forms, which were employed in my experiment, are capable of conducting lithium, or not. Some myriophyllum plants, firmly fixed by means of cotton wool, were inserted into a hole bored in the cork of a small glass-vessel which was filled with 2.5X Li NO3 solution, and kept under water during the experiment. I examined them on INJECTION' KXlM•:Kl^[^;^T.s ox plants. 11 the following day, and confirmed that lithium was present through- out the long stem, but less in the leaves. We may infer from this demonstration, that in the case of injection, lithium can be more or less conducted through the body of the submerged water-plant, but the quantity of lithium conducted is so minute, that I could not detect it spectroscopically. As to the fact, that lithium, being conducted, was not stored up in the submerged body, but was lost from the injected part in course of time, we may imagine that lithium in part exudes out of the water-pores of the leaf or other parts into the surrounding water, and in part diffuses into the water from the wound made by the injecting needle. ■ 2. Upafaei) conduction and downward conduction. We should not leave the fact unnoticed that the injected solution was conducted not only in an upward direction but also more or less downward, which we see in almost all of the experi- mental data. The upward conduction, as a rule, was observ- ed most conspicuously, especially in the case of transpiring organs I observed that lithium travelled up to a distance of 337 cm. in Trichosanthes cucumeroidcs (exp. 15) and 203 cm. in Cornus con- troversa (exp. 5). Here I may tabulate the comparison of upward and downward conduction, taking examples obtained in my ex- periments, as follows. 12 ART. 6. — Y. YENDO Plant name No. of exp. Upward con- duction in cm. Downward con- duction in cm. Fatsia japonlca 34.5 2 Cbmus controversa 5 203 — Ccmna lutea 18 57 68.5 Setaria sji. 19 133 2 Allium fistulosum 20 8 Rhoco discolor 20 9.5 2.5 i) }f 21 10 1.5 Gelosia cristata 44 5 Euphorbia heterophylla 57 2 CarcK macroeepJiala 26 49 12 Cosmos bipinnatus 7 89 80 Asdepias curassavica 22 10 Tetragonia expansa 8 21 19 plus. Cotyledon malacophylla 9 11 1.3 Platanus orientalis 35 5 Rhaphanus sativus 5 2.5 Vicia Faba 8 6 Chlystegia soldanella 13 21 37 Tricliosanthes cucumeroidcs 10 337 10 Fueraria Thunbergiana 17 197 7 Jussiœa rejyens 30.5 16 Cereus serpentinus 61 31 7 Lycopodium cerunum 40 9.5 1.5 Lycopodium clavatum 41 2 25 Equisetum arvense 42 18 7 Oortinellus edodes 45 ß.ö 1 IXJKCriOX EXPERIMENTS OX IT.AM.S. 13 As we see in the above table, upward conduction exceeds the downward in many instances, but sometimes a vigorous downward conduction^) occurs. In Caima lutea (exp. 18), I observed that the injected hthium not only descended through the stem down to its base but was also conducted into all leaves which attach to every node. A marked downward conduction also occurred when injection was made in the leaf, flower-stalk, fruit, etc. SO^Cu was injected into the leaf-sheath of AlUmn fistulosum and penetrated down to the rhizome (exp. 62). In Narcissus painjraceiis (exp. 23, 24), lithium, when injected into the leaf or scape, was found in the rhizome and most coats of the bulb, moreover it was also conducted up to other leaves and the scape. In Citrus nohilis (exp. 3, PI. I, Fig. 2), lithium, when injected into the fruit, was conducted backwards through the fruit- stalk and branch and into 4 leaves. When injection was made into a long rhizome or stolon, I often observed conduction in both directions, forwards and back- wards, (exp. 13, Calystegia soidanella ; exp. 14, Lactuca repens ; exp, 36, Carex macrocejyhala ; exp. 42, Equisetum arven^e). Indeed, I observed in Carex pumila (exp. 25) that lithium travelled through the rhizome fore and backwards, passing through the erect shoots one by one. In the experiments with leafless deciduous trees in winter or early spring I could also confirm that both conductions occurred, the upward being more vigorous than the downward one. 1) ScHECHXEK (16) mentioned that a vigorous dowmvard current is due to the inversion of the osmotic ratio and according to Boehii's (17) view it occurs ^^■hen the earth is relatively dry and the sap-conducting vessels filled -with ^vater. 14 ART. 6. Y. YENDO Plant name No. of exp. Upward conduc- tion in cm. Downward con- duction in cm. Hamamelis japonica 50 30 20 Cornus oßcinalis 44 17 Fagus sylvaüca 51 15 9 plus. Calycanthus praecox 52 6 2 Cornus controversa 53 32 15 )f 54 G 7 24 }} 55 74 10 }) 56 19 }} 57 57 25 Prunus mutab'dls 58 32 1.5 3. Transverse coxductiox. The injected solution was conducted not only longitudinally but also more or less transversely (Gaunersdoefee, 6). In Cornus €ontroveisa (exp. 54, 55) lithium was found in the wood of the injected side of the stem considerably wider and deeper than the dimension of the wound in which the injection was made, but did not spread so much as to reach the opposite side. In the case of a smallej' stem, lithium was present through- out, but the amount of lithium in the injected side always exceeded that in the other parts. We learn roughly from these facts that transverse conduc- tion takes place moderately through the lignified wood. I also observed less transverse conduction in plants whose vascular bundles are arranged separately, being interjected by soft parenchymatous cells. An apparent transverse conduction often occurred in plants which have transverse connections amons: the long-itudinal vascular INJECTION EXrERIMENTS OX PLANTS. 15 bundles ; in this case, the transverse connections may effect that conduction, for example, in the parallel-veined leaf of Rlioeo discolor, provided with many delicate transverse connections among the longitudinal veins, I observed that lithium was conducted transversely to the margin (exp. 20, 21, PI. I, Fig. 4. A, B). 3. Speed of conduction of the injected solution. A rough idea about the rapidity of movement of the injected substance was arrived at in my experiments. It was ascertained that the speed varied according to the substance employed ; lithium was carried most rapidly, but SO^Cu and eosin less rapidly, and aniline dyes the least. Indeed, aniline violet was conducted so slowly that it was left just as it was in the injected part, even for some hours, but was then slow^ly conducted up- wards and at last so far that the extension was no longer detectable. After some comparative experiments, I confirmed that the rate of conduction of several aniline dyes (Oveeton, 8), varies according to their kind, i.e. a basic aniline dye (for example dahlia- violet), was less conductible than an acid one (for example <3ongo-red or indigocarmine). The relative rapidity of movement, however, depends on the concentration of the substance employed, because the proportion of osmosis and diffusion may be altered by varied concentration of a solution (Gaunersdoefer, 6). To ascertain this fact, I demon- strated with the leaf of Fatsia japonlca, applying LiNOg (10^, 5^, 1%) solutions. After a certain time, the leaves were removed from the stem, several deep incisions were made with a knife along the petiole, in order to break the continuity of conduction, then examined spectroscopically. 16 ART. 6. — Y. YEXDO : 0.03 c. cn<. of the solutions ^^'as injected in each case. Date of injection and remarks ^--^iNOa Duration^-v,^^ of exp. ^~""\ 1^ 5^ lOX Nov. 30. A.M. 10^ 37m-43m. Fine weatlier, somewhat windy Humid. = 48^-42" Temp. = W-l'o" -j hour I'^cm. 2.0,,., 2.3,„, Dec. 1. A.M. 9° 20m-30m. Fine weather Humid. =61°-6j° Temp. = ll'-12' »j-hs 20™^ 3.6en,. ö.4cn,. O.Ocn. Dec. 6. A.M. 10° 20ra-24in. Fine weather Humid. = 75°-55' Temp. = 14°-18° 5"^ 40™^ 2.8em. 4.5e,„. G.Oen. The distances are those of upward con- duction. The above table shows that a 10^ solution has been con- ducted to the farthest distance, while a 6% solution was not so conspicuous, and a 1% solution the least. 5. CoNDrCTING PATH OF IXJECTED SOLUTIONS. Applying the aniline violet solution to Vicia Faha, I observed that the dye was conducted only through the xylem portion, i.e. the lignified walls of the xylem elements were deeply coloured and contents of the xylem parenchyma cells more or less coloured (PI. II, Fig. 2). In the rhizome of Menianthes trifoliata, I also observed that eosin was conducted most easily through xylem por- tions, especially vessels (PL II, Fig. 3). However, in most cases, phloem elements as well as parenchymatous tissues were penetrated in various degrees by injected substances. I observed in the ex- IXJKCTIOX KXl'KUnfKNTS OX IT.AXTX. 17 periments on Cereus ssrpentinus (exp. 01) and 'Tr achy carpus cxceha (exp. 60, PL II, Fig. 4), that SO^Cu was more abundantly present in the phloem than in the xylem. In Lycopodium cernuum the central cylinder, more especially the xyleni parenchyma cells which enclose the vessel, contained much SO.Cu, and the cortical sclerotic cells more or less of it. In the experiments with LiNO, solutions I ascertained that litliium was chiefly found not only in the wood, but also more or less in the l^ark and pith. In short, we may infer from these facts that the injected solution is conducted most easily through the xylem elements, but also penetrates the phloem parts and other tissues in some degree. The rate of penetration of course varies according to the substance. 6. IjS'JECTION method as a MEAXS for TnA('lX(r THE COURSE OF A VASCULAR BUNDLE Many physiologists have investigated the course of the trans- piration-current by placing cut branches in the solution of a colouring matter (Pfeffee. 11) or by means of precipitates (Dixon, JoLY, 20). As I had opportunity to observe this phenomenon, some instances may be mentioned. A small quantity (0.03-0,15 c.cm.) of 5% SO^Cu solution was injected into the petiole of Falsi a japonica on the upper side and examined after a certain time. Tracing the vascular bundle which contains SO.Cu (PL II, Fig. 1, C), I found that it divides into some veinlets {!>') at the basal part of the blade, and they assemble as if forming a plexus (c') with veinlets of other vascular bundles, but again separate from one another {cV) and then run into the blade as palmate veins, some grouping together where they arrange themselves in horse-shoe shape (e') as seen in cross sectitn. It is 18 ART. 6. — Y. YENDO : interesting that the bundles containmg SO4CU are regularly situated on the upper side of the horse-shoe shaped vascular bundle groups in all tlie palmate veins. Many veinlets are given off from the palmate veins and terminate in the parenchymatic mesophyll, forming a network which is often blackened, due to the action of SO.Cu (PI. II, Fig. 1, D). The lithium injection was made into the pulvinus of a leaf of Magnolia hypoleuca (exp. 4, PI. I, Fig. 3). The point of injection was situated on one side of the pulvinus. The result obtained was that lithium was found in only one half of the blade. This shows that a vascular bundle in the pulvinus runs through the midrib and branches off as veins and veinlets only on one side of the blade, but not on the other. A similar case occurred in the ex- periment with the pinnate leaf of Aspidium lacsrum (exp. 39, PI. I, Fig. G), whose vascular bundles are separated one from another by the interjacent ground tissue of the petiole. I observed also in the experiments on woody plants that lithium was detectable only in the branches belonging to the injected side, but not at all in those of the opposite side. Thus, by the above moans, the course of vascular bundles can be determined. X. Summary. 1. The rate of conduction of an injection varies according to tlie nature of the substance ; among the substances used in my experiments, lithium nitrate is most easily conducted, SOiCu and eosin less easily, aniline violet the least. 2. The injected substances are mostly carried to that part where the transpiration is going on naost actively. a. Lithium, when injected into the stem, is found more in ]N.ii:( rioN K.\i'i;uiMi;Nrx on plants. 19 louves than in iullorescences and young parts. //. lithium, when injected hito a part of the cotyledon, rhizome, bulb or root, is carried up to the shoot. c. In some water plants, e.g. Myriophyllum, injected hthium, is found abundantly in the œrial parts, but very little or none at all in the submerged parts where the injection was made. 3. The injection experiments performed with i)lants cultivated in the moist chamber show a ^•ery feeble conduction of the sub- stance injected. 4. Conduction of lithium injected into deciduous trees during winter is more or less perceptible, but is much less manifest than in leafy plants. 5. Upward conduction is always conspicuous ; downward con- duction generally less so, though sometimes very manifest ; while transverse conduction is very feeble. 6. Injection solutions are conducted mainly through the xylem <;lements, and also through the phloem parts, though other tissues permit conduction in some degree. 7. The speed of conduction of an injected solution varies according to its concentration. 8. Certain fungi {Omphalia sp., Caprmu>< sp., Corfinellus (idocles) indicate conduction of injected lithium. 0. Algoö [Sargassiun Ilorneri, Ecklonia cava, Bryosjns imhnosa, etc.) scarcely conduct injected substances, but mere diffusion of these may occur, and an injected solution passes out of the injection wound some time after into the sea-water. 10. Vascular bundles are traceable by means of the injection method. 20 Ar.ï. 6. — Y. YEXDO : XI. Experimental data.^) Experiments with LiNOg solutions. (A) Vascular plants Land-plants Dicotyledons Exp. 1-17. Monocotyledons Exp. 18-26. Pteridophytes Exp. 39-42. Water-and Marsh-plants Dicotyledons Exp. 27-35. Monocotyledons Exp. 36-38. (B) Thallophytes Exp. 43-47. Experiments in moist chamber . Exp. 48-49. Experiments on some deciduous trees in winter and early spring Exp. 50—58. Experiments with SO^Cu solutions Exp. 59-64. Experiments with colouring matter solutions Exp. 65-76. Injection into parenchymatous tissue Exp. 77-82. 1) Tlioiigh I, experimented with 73 species belonging to 49 families of A'ascialar plants and 9 species of thalloiihytes, I have described here only about one third of my exi^eriments, selecting the moro significant cases IXJKC rio.V KXl'KlllMKN'PS OX IT-AXTS. 21 1 n'iS c ^0 0 ■3 •r-t £ (N d a 3 .3 •T. J-3 is fl 0 0 a 0 M ft a ft C3 a 1 to _ . 'S a 'a é'il 0 n ^ S p 0 g a OJ ä a ^ ^ 's M a m û ft 'S .'s" 0 1 g ft 0 tn 'S te a 1 to -i II 10 CO ■Jl 1 03 p 0 tn fH es .9 'G -^ -^ ft =^ m: g -5 a " aj -Ö tS "" ;^ •■^ 54-i 0) ft 0 e: 0 S 1 ft if 2 0 «a S Cl ft 0 s 2 2 © ,a iß 3 S-l CM ft 0 _a a ft aT -^ CD 0) a .2 to > a 'S 0 CO '^ n ^ 1 i .s ft . ^ a & ^ ? a -2 ÎD 'S OJ • i-3 •I-t •i-l • f-î •r-î kI H^ h-l hJ Kl h1 *-5 ^^ hh" D ^ 05 1^ Pk ^ " jj S «*H t3 0 2 a d 4-1 =0 a g« tH 0 0 r— il ■^ ,2 d "o CO 'S a^ in 0 a to § 1 a a _a a a 1 S T. 0 3 S C5 1 '" -a H «+H 0 CCI rH 00 t^ 00 r^ [Parts ex amined after injection .2 1 -2 1 i-i ^ d ai fi g d fi § 1 1 Cli P4 eu Plant name and famUy name AH «£ W U ü 1 ^ ^ e S « c3 2: 0 g tH !M M 22 APvT. 6. — Y. YENDO ^ -4^ r- (C o '-' (^ 5 ois 1c '? ^ P 03 O a'-gf a t! '^ 'S a s s « a 1 O 5 P CO fl co 2 S "S ^ t: "Ôj 6 îl^ co a) 1 O a a -5? ? o '^^ S"« î; c 'o ■ « o • /3 o c2 fi & '►3'""" fi 3 .1 ° "m a; ■" "Pi S o co a'H (N c f^l? fl*' ill "^ Ig -s ^ 13. p: o S ^ .a -a 5 'S o X to s •M J O _ tß £ ß ^ -S CO 75 s '-i o 5 a 0) g.s co , £ fl ^^ . çu s ® ■'^ CO CD r— t & co co « M 1-:! 1-^ '^ ^ 3 13" S h1 p a _j^ ce a cß tu i 'S n 's 1 co r-l ■S •S, H " ^-c tD^ • O 5 œ >> a co a 3 co a O j' §1 a " te 1 te I S in 'co o co u 0 S 3* SE . ä ^8 fi '". lO es' o ce CO tH «N rH co ■^ x' lO Parts ex amined after injection "S ft a o 3 ■S 1 1 a s. J es ■ - £ a te s CO •S îj-i si -^ .s a co g a- a !>>r^ rs g ^ H « g o ."S _£ -= O c o «^ uo r' S.S.. ô ^ CO u -^ -^^ o Ö ri "3 tH " O Ö cT p . 1 M ^ j^ uratio of ixperi ment ^ë o3 -= (M ^ ti lO 1 O lO I CO ß " r-l (N (N (M o" (-H* o J^. H '-' 03 r^ ^ ^ '"' 0} 3 ••-1 a ..s il Û.1 a -SM 3 8 ^ ce S i . •§ ifi ;li O T II a i o r^ ?;■-! Ml m - 'ih INJECTION EXPERIMENTS ON PliANT.S. 23 87cm. giving off G branches at its lower part. ' Li. was present down to 80cm. in the main stem and more or less in all branches. Li. M'as present down to over 19cm lii. was present in one leaf, absent in other leaves and young shoots. llcni. bigh. Li. was present down to 1.3cm. in the inj. side and in 3 leaves. forks at a distance of 3.5cm. below the inj. pt. Li. was present down to 3cm. Li. was present throughout it and in all leaves. Li. was present in the stalks and in both inner and outer involucres, but absent in ray flowe.rs as well as tube flowers. Li. was present throughout. Li. vi-as present abundantly in all leaves ; in nerves as well as mesophyll. Li. was present a little. Li. was present up to the ' termination in the inj. side. Li. was present in all leaves and bracts which attached to (the inj. side of the stem. Li. was present in the calyx, absent in petals, stamens, pistils. Li. was present throughout the inj. side, and abundant- ly in the central cylinder. Li. was present in several leaves, but little in its blade. Li. was present a little in the scape, absent in the involucre and flowers.^ Li. was absent. 03 00 g ÙU 21cm. long; 14 nodes, attaches to the main stem and axil- Lary shoots. Hem. high including the height of a spike. many on the spike. 9cm. long. (a) flowering one. (;3) unflowering one. stem infloresc. stem leaf termi- nal bud stem leaf flower root leaves infloresc. a to 1 CQ Q ". in îrt 25 ^ o «? S \o O y ^ â g •^ o CO 1 r-( (M r-l "t "^ 1 e Ä O "8 M o '^. ft 2 ■S o (D to IH O ^ 3 S 5 <=> o «^ (M o . cr: a N rt »*i ^. Cosmos hipinnntus. Cav. (CoMPOSIT.) Tetragonia expansa Ait. (Aizoac.) Cotyledon malacophyla Pall. v. japoniea Fk. et Sav. (Crassulac.) Taraxacum officinale WiGG. (Composit.) t^ OC' 05 1 o I-H 24 ART. 6. Y. YEXDO TD -2 1 o p, it Ci ft Dimension, number, position and results Li. was present to 6cm. below the cotyledon, .absent in all side-roots. Li. was present back to 3.5cm. 27cm. long, stands at a distance 2.5cm. from the inj. pt. on the rhizome. Li. was i^resent throughout the stem, abundant in all leaves and a little in the calyx of several flowers but absent in the corolla. i5 'S CO Li. was present within a distance of about 2.5cm. around the inj. pt. Li. was jjreseut in all parts except the young leaves and buds. Li. was absent in all jDarts. do. Li. was present evidently in those attaching to the shoot (a). Li, was present abundantly. Li. was i^resent tbroughout. Li. was present in the sti- pules, petioles, and blades of all leaves, absent in the apical bxad. Li. was i^resent throughout, Li. was present in the stem, abundant in leaves, absent in a flower-bud. Li. was absent. u a 6 .il fl o lft ■ iH ft fè it ^'S" . CO ® .fl ® Cu.fl.fl „„-^.SPöO^ .a "t! .£p fl -^ .a ? a.fl-g a pp: a 2 2 ..§2 i2 2 t£,=« Ç 2iico ec=ü ^ « ft..-^ Ca ft_. ii ^ 2- ^ pt.; 7cni. high, attaches to the basal I^art of those shoots. 03 S _03 fl lO ä 10cm. 11cm. high, stands at the termination of the rhizome. Parts ex- amined after injection bulb shoot root coty- ledon stem leaf root rhizome shoot root Part injected 1 coty- ledon rhizome Quantity nijected v= o i '« 3 q " o - 1 i. ^ 1 ^ ifi 5 =2 hi ■ Dura- tion of experi- ment _ CO rH rH O 2. M ri< Date of injection - S s o rH CO ^ o" 03 ^■ ^ 2 Ph ' ■S >B 00 CO rH c O CO g g „ Ë ■z o 00 'iH ri c- 03 <=■ - CM Plant name and family name S 5 7 Vicia Faba. v. equina Pees. (Legnmi- noc.) g 1 M O ^ d "tj ft rH rH rH CO IN.IECTIO.V EXI'KIUMK.XTS OX I'L.VNTS. or. Li. was present back to 2cm. 1 a § fl 3 ft rH O 05 Li. was present down to 4cm. Li. was present in the vici- nity of the inj. pt. Li. was present in all leaves, but absent in the rhizome and shoots. do. Li. was present up to 337cm. Li. was present in the jjetioles and blades of 45 leaves. Li. was present, except the upper ones. Li. was present up to 11cm. in it and continued to a branch (lOcm. long) bear- ing many inflorescences. Li. was present in the pe- tioles and midribs of all leaves. Li. was absent even on the branch which attached at the 2nd node. forks at a distance of 16cm. from the inj. pt. (a) 68cm. long, with 8 leaves and many young underground shoots. (j8) 31cra. long, with 6 leaves and many underground shoots. 383c„,. long. 58 in number Avith a terminal bud. 98cin. long ; 16 nodes. a leaf attaches to the 1st node. 4 leaves on the branch at the 2nd node. rhizome stem leaf tendril stem leaf infloresc. rhizome stem a 1o lO '« S o B « rH "3 O \o o S §1 to in C^ 1 (M •S O ' 3 5 ,n 00 CO r-l o' Ä o a (M o Lacluca 7'epens Maxim. (Composit.) Tricho- santhes cu- cumeroides Maxim. (Cucurb.) Dioscorea Tokoro Makino. (Dios- coreae.) i-i 1^ I-H 26 ART. 6. Y. YEXDO ft^ fl ^ 03 Ö 3 's M n ^ 'S 2 tb 0 d Ö o rS 'S g- a rH ft § "n sn hfl M fih TJ •s r! *" * OJ fl O fl s 9 en S ft 03 a •I-l "S a 1 S 03 es a 1 03 •S 0) 03 O) ft 'S 2 ft eä r- to t> „. •-C ft§ ftœtfl h1 p4 HÎ Hl h1 hI rë ^ o O •73 tS' 1 a o g» § (0 (D (S "es o ^2 1 rfl ■fl • <-( 1—1 © g m g (B ® ^lll J Ü 1-^ LO 60 a 01 c3 0=! tM m il fi a n •i-i i •a . s a 00 CO "r? 4 S 02 I-l 1 fl (N i 1 6 'S" "S d . œ B^ intg 00 IH (M Ä-n e 1 u 03 1 Parts e amine after injectio ta 1 a -2 'S .2 ..S 'V a a Cj S œ (U f..l IQ CQ ^.rn « B ^ f= ^ O =? \o o c; ^ -^ 2 in ?5 in o è ° '„^ fl m ^ 5 5 •o S œ "S Sal rH S o Dat inje l-^S O P4 .- —« Ö Plant name anc family name «'S S H a i -3 4§ ri 0 ^1 0^^ eu t» 00 1^ . 01 M . Ö i> to S; S a ^ u œ 5 câ IS »^Tj h1 H ® cl ^ s ^ '53 _ (D -M ^ CD O tS ij 5 2 ..s 3 5 13 Ö ^ ö.Ü g« S-b g. y & ft .-, s & 03 CQ 0 -U & '^ ä ^ M =3 ^ ^' S . s ft . o,û te W) r,n o 60 r,n 0 a 0 0 o o ' '■■* e S E S 05 <Ä lO ■>* CO ox Ol S - -^ W cô ^ M .9 " ft s 60"-=. -71 fl O a "S ^ ^ Ci ^ G ^ ^ c3 — ' 3' S- ci Cd — e- .2 W '^ S «Jri^O fl o o ci-. s ce .2 H 2 M S S ai 28 AlîT. 0. — Y. YENIX) is PI n 'i f M ^ S ^^ to 02 5 ^^ ■i-î "^ h! as fl^ r« 01 o ö t>. Ö 0 a f- 0* 0 o 1 rij œ 1-H Cï ^ c3 es is g es > O g a a -^ fl .g ^ 0) »"^ ■^ hI "îH -t- ä M ■ "^S ""1 o'g ij Oh n (D 3 n o ^^ M s g Sfl6 ^ to ft s ai ? œ © eu '3 d d «4-1 »H fl . t: C3 o ft cs as 5 .H 3 ft S ^ o r fl fl o •^ O o 0) ft S " S ^ 3^ to ;a o.a .,5 o 5 J5 03 r^ = Çd ce fO X -Ç 0) CO -i n n't A- cS H "3 m ^1 _ qj H P Cl a 2 fi-2«a es .s, «.a § sa a o u-i ft ^ o X ^ S «= S =-i S "^ h1 5 rH I Il «'S" « c w c; INJKCTION' K.XrKUlMKNT.S ON' I'r.ANT.-; 20 Li. was present down to 28c ra. in the rhi- zome. {(t) 2 shoots are standing at a dist. of 13cm. from the inj. pt. Li. was present in theso shoots. (;3) 1 shoot stands at a dist. of 28cra from the inj. pt. Li. was present in it. 10cm. to the first node, Li. was present just to the node. 7.5cm. long. Li. -was present abundantly down to the base. d 2 p. M Li. was present forwards to 7cm. «T'ttd continued to a shoot. Li. was present, especially in leaves. Li. was absent. Li. was present in the rhizome and one shoot ; present also in some leaves and bracts on the spike of the shoot. Li. was absent in the rhi- zome as well as the shoot. lii. was present abundantly. Browning occurred at the margin. Li. was present in all parts of the leaves. Li. was present up to 6cm. in the. inj. side. Li. was present in some of them. 73cin. long, bears two shoots. {a) stands at a dist. of 7cm. from the inj. pt. (3) stands at; the end of the rhizome. forks at a distance of 2cm from the inj. pt. (a-) 16cm. long, gives off 3 shoots at its extremity. iß) 16cm long, gives off a shoot at its extremity. Qb a c d ,2 Ö -• d e-3 y> a shoot is growing at its extremity. attaches to several nodes. rhizome shoot rhizome 03 0) a ^ .2 g 'S CD i .s 13 -> "3 1 1 a 1 ft aï m ce ci 1 ft © ft ce i CO a 73 ce (B œ > to [o 4^ • (H • 1-1 A .3 "ce Hl 3 l-q 3 ^ ^ 3 3 M i-:i a "' 3"^ 'g .2 1 M CD •iH (M to «3 03 1 03 fi il s ri a 1 Ü D a 1 ce CD 'S 3 O 'S O .2 a 1 a r- S g E g tp r2 5 "ce 2 4^ 4 S in |S 00 o ri t- ^ rd 00 t- es Parts ex- amined after injection ,2 ® .2 's g ft o ® r % ft » S 03 a •t-t r-l 'S 1 'S OJ 2 S « ee «J a Pi ^li ê'sl § a^ Sj ■. 1 8 X 4> Ö 2 ft •r. % CO a to i a -^ o t rS 1 'S «a ft .2 03 1 ë 1 g tu o S vi a -^ ig .S ■^ 1 'S ^ HS 'S 03 ^ ^ 1 : :: 03 03 03 •(H "o 0) ft 03 ^ o 1 â M a ^ 2 II es -^ '^ ft "3 's S-l o ® ■ P 's i? • f-i •1-4 •I-I • f-* •l-( Hi Hi Hi H^ h! ^ i^ Hl "l % :o ,g lO -a M (E 1 Pi fl • rH 1 a <0 1 1 CQ œ o 1 03 (D M 1 i i Î J ft o S a 's 1 S CO c? GO i^ O P! 2 t- r-l rH c3 rH 03 03 rH ci •3 ^-1 u I O a - "S a 03 cS -ö 'ö 10 1^1 logg ^-1 ci s. *^ à n à « S âO p o g \= 2 Hl in o ^ CO ^° 9 in rH '^ O Ö r-l Ö O Ö 13 rH Ö Ï Ï ^ (N Ol S l-H ^- g -• S' rH M ^ 5 ^ o f^' ^ 3 S » ■""" PI •û s ^^ ■y. i3 ^ a! as ô ^ . f 'S Ö" s /H 'S h\ s Ä rS. .^ a3 '^ i o C o a o ce 2 'S o SP -s p a> "3 ^ ■*^ c; p g -s sa *^ c ^ r-l s a (M M • ^A o O 03 1 ce r2 1 ce Pi •(H o S -5 .'S 2 S o -j: 1 a ei Ol (M ct-l o «D f» c; O g PI c 1 ce te .2 pH 'S 3 PI S ft 1 to 1 g te ce ce 'S . "S a: i| il ■^ 2 ^ _ .2 ta ci œ 'S Ol 1 x TT. § & a; (S > § ^ 1 '3 § p _2 te Ol ce PI 1 a Ü ^ ji SM fl £ Ö c; A r^a Pl -y o • ,^ .^ •fH ■i4 .«• -^ .«• "^ '^ "^ 'r^ ''fi- & 1 H^ hi H-l hI h5 h! hi p 1 >^ >^ © r- ! ^ c3 O r3 i a a 'X CM ce PI CO ce ß ■^ 1 II o 9 i^ CD PI •fH a "^ 5 SE CO o 05 03 O o 1 "S Œl "S .2 "«s ^ (M •?. . ^'? £ 5 (M Ä «3 c; Parts ex- amined after injection a -2 'S 1 o i PI 1 i ce r- — ' -ri O 1^ -< o' I-l Xi •A es cî a; œ — ' 2 la i 2= 3 I II 1 .3" i » ft Oh- CO = o S s 1 s S 1 § h1 o ;3 § 1-^ o lO g g h1 ö •n CM rH 1 (M il* C5 (M 1 (M (M ê (M 1 00 rH 5 o ^ ^- - 2 3" ë < rH "S g 5 o O eo Lycopo- cernuum L. (Lycopod.) agi 5 .S 1 h; 'S •2 ^ o ^ s . Eh o ^ I 01 -* •* 34 AE,T. 6. — Y. YEÎ^1>0 05 - fi fi al c3 S &•-' ^ CO 1 a o ^ re fl fi| CD cS 1 'm .1" 2 fi % 1 3 j:^ CD O Î-1 2 ^.a ri § 03 y^ r^i a Tj rH .2 • ^".3 CO 01 CD to g 2 ÖD e: o .^1 1.| \o o s ^ >Z ai m !^ o \o O \o O' in ^ 5 CO o ^? g CO" in W o a5 (M M i ^ §•3 . fi Ö fi o ^ o h! d 1 ^ Duration of .experi- ment •3 (M ■S §» Z CO in ^ ^ O (M O 1 ft a fi e fl m o c fi o eu ä o o "S.S. fi. a .= o o -tj o -fi CM 3 (1^ o rs CO t s fi ft O ^ — , Plant name and family name s ^'0 « fi Ir' p a Ö .S ^ rfi PC» a, ^ •S 1 M a ë 3 ^1 CS CO 2 •'S "'S ■ — -— ^ ft ci /^ «s: ^ Hi ^ lO «3 t^ 00 ^ "^ « ■«JH •* ■* ^ INJECTIOX EXIMORIMENTS OS I'r.ANT- 35 > DO P 5 ai a.o a 'S OD e: CO a p" d _2 -^ an CiCtS H 9j « 2 5 H .^ .2 'S »^ 5 .a a.2 ' H ■* .a !_q _o (B CO o^ U CO a; +3 ^ 35 is5^ > "^ 0) o *^ -a fl o 2 CO ci ~ S .'S P. eg tc (3 O >-§ ® a rt » m "" >-i *^ -«^ Cj t2 CO , o ~ -do ,û o Ü c; ■" O ce-: % o tc S ^ g 2 ö o ^ O c3 CO ^^ 3 "* ,TI tI i? TT ^ CO A Tl ,_H O ei PI 0 (M y E o .iU CO iï A Ö « 0 v.. O 1-^ 3< t 6 ^ •5 '^ 2 e 'PM S ft ,h:i f> ç-fM'3 O a 2 «2 W 36 AET. 6. YEXDO , u ri fl o .«■ a a> r! a> 0 ^ -^0 > o o _^ o 'S 5 'S ci a-1? g g ce 1 CO '^ rt OJ Pi -1 ^ '^ S g •a 1 5 1 [S 1 o in rH O 1 g 2 S ä -'^ ft -? f=i ^ (M CS !/! a 1 o CO P -Ä ft O a5 ■^ 'ce '5' sa h:; ïï 1-1 00 r-l -2 ft ft o R d ci cï 'S •rH a CO ft nt in several h attached to but never chlets. a to stem, more sent in the d pith on the 1 1 1 CM O g CO fl a •rH ce "S fl S a 'S 'S c "S 0) and also prese branches whic the inj. side, in small bran fl PI a> d TS 02 Ph ft (S m te ai H ft 2 ft m et c2 ft a -. -a m -S -Û -Q r:; IT3 fl fl 1 'ce 0) ,d g  'S et f: p .a ^ ?* ^^ is 1? 'fl ^ 'a '% r^ P 1-4 H^ M hI h\ t-4 M a in Sä O t- 0 g« ßg go a es CD 1 té "S £ ce -2 o n •r-( M 0) ts o qd ce a; 1 •g i cS F o 1 's c •>*i d^ R H ,Q rfl -ij g a o ce Ö u o tH tH iH <» (N "^ E Parts ex- amined after injection rd ,cl O r^ t~ 0 ,S M 1 a -2 5' pQ ^ Ifl ,a ft ,2 Fart injected o ci a C3 a i .i 05 ^C •f-t pQ ,J3 ^a .a «.s cc •fH f^rr- B p " 1 ^ S 0 to \n O J^ V - o 2 O p-i o S r-l o " i m ^^^ 00 o in' b ?; Ï5 § '^ ^r f- G? -a (-1 o ^^ c H^ d ^ d g fl m ë uratio of xperi ment ^1 C<1 ■a ■a in (N iH (N CO r^ l-H Q ^ (N 0 O O s: O S r-( o (N S in o' Cl p. lb i is CT "S P4 r-l i— ! ft o ft 8 Ci C4-f C r^ t= s "^ i ^ e 00 'co riant name an family name ^ .H •g Ö ^ 6 i 2 8 •rH ft o «H A r-t fM M •<* ''< d in m >o in INJECTION EXPERIMENTS ON I'l.ANTS. 37 ci .J a ,a tc -.Mm P 5) ?^.a 2Ü -♦-» H^ Üß CD -" -M :t^'n §•8 "? •i-t &"5 ro-^ - H œ r— •- a /^ s ; »3 CQ eS -^ h^ PI M CX) sc Cß tJ3 ^ ^ O * e ?î ca 1^ -. »g fi j^«.0 M o f> ce QJ ,^ 3 0 il ^ "er fi O is . ■S3 g o o B &iCO •S =^ ^ g ® ^ 2 ri a S — PI > fl â 9'^ -fl-ö is ^ W H n'a , -^ CD q3 p:^ 02 S "^ « rt CO ci .a Ci a CO '5 PI sas _ o PI ÇO CO ^ ce es .. p;.'S-3 en oj CO aj o œ c3 cs a n CO -w >H g » a^^ s ,2 o ö fl "z d ^. œ o r '^ .g , p £ 2 s a ® ^ if? af^û C3tt3 OJ, od .9 CO CS m fs ".g S =0 5S o o ^ § g a '^ S ft œ !in ■a ÖD 's § a P ^ ^ ^ T^ tu ^ ÖI :^ ;/.' P ^ ß à-=: CO --H H vo .g ^-= .9 lO o Ö H = ja uij CO *^ o a * O M -£3 (D a ^ PI .s o o B 0 ^ ^ fi-^^a o.o ft. g o P4 flK 2 '3 a •= « s s w - ÖD ft -^ E^ <: i^ i .4 Q 9« s-ssa 9« 1^ s s o .so ; ;rq INJECTION ]:XPERIMENTS ON PEANTS. 41 Aniline violet was present down to 2cm. It was jjresent down to 3cm. Eosin was slightly present about the inj. part and contenta of that part were disturbed in their arrange- ment. Aniline violet was present in all tissues about the inj. part, but did not attain far; the wall of vessels was well tinged. Aniline violet was present up to 2cra. in the paren- chyma and vascular bundles ; epidermal cells were well tinged without any appearance of injury, especially protoplasma, nucleus, nucleolus etc. were distinctly tinged. Aniline violet was much decolourized and only slighlly visible. In the inj. part, chlorophyll-grains, and other contents were tinged and disturbed in their arrangement, the cell wall near the inj. pt. also slightly tinged. Dahlia violet was detectable up to 4 cm. Several tissues were tinged at the inj. part. 4« ai 1 r* Ci Mm "œ _ci 1q .» to -M ci .2 'S 0.2% Aniline violet O.UGc.cm. c?.at3 s CN r= ■= "-O O S 'S 1 =3^ 5 \o'.ä o • c fi ^ 5 r-l « ci r-l M CO iH â 1 i §i 5 iö ^ o' ca i-i r-l o ^. J3 IH l-H "5 ■i < S .= S •3 {h 1 N CM .2 e ^ il il Raphanus sativus L. (Ciucifer.) fistulosum (Liliac.) Bryopsis pulmosa KÜTZ. (Algae) japonica Done'; ET Planc. (Araliac. ) o rH L'- 0-1 L- ^2 S 42 ART. 6. — Y. YEXDO CO d II it Is II d 1 d a; SS CO ^o (—1 d £: o 1 CO ^ o Li, was present down to 14cm. Aniline violet was present down to 1.5em. Upwards CO 1 Indigocarmine was detectable slightly in several tissues about the inj. part, some elements of the bast fiber thinly coloured. Congo-red was present up to 12cm. in the xylem of a few vascular bundles. g _c 1 CO g ft Li. was present Tip to 8cm. in the main stem, not at all in other organs. Aniline violet was present up to 6cra. in the section through the stem at the inj. pt. The dye was found in most parts of pith and xylem portions, but not at all in the oxxter tissues. In the section l-2cm. above the inj. pt. the dye was found in most xylem portions, but absent in the pith. In the section at Sera, above the inj. pt. it was found only in few xylem portions. Û , . S c fl œ £ ■ ft ^ c œ c K g ft ^ ce Z es-* .^ ri Ci TU .^O ^O ^. O ^^ 3 h1 g œ ë o (^8 v,4 o Î3 . ÇD ^ o /-. ^^ «g /h 0) S 15 s "K ^ •*-> ■Ä -t-' ■r; "S t3 M t; œ ft«-. O C4-I ° «H O O o o o CO r ^ g -o s ^ Ö X= O \ ^ !^ O Ô^ ^ o" O »fH o i-î Ö ^ S C Ö -^ j ". ^ M ï o- -^ •â q» •S cô •3 «3 J. CM 1 Lt. 1 o rA (M (M ^ j2 ^ E +^ O p K lO S ^ SO (5 ^ ^ If <*i o' o ^ g" O i—t ^ Ph' e « e ^r '*^ et4 — •ïs .S . «^ t Su ^ ^£ OQ o O cS O M ='■ S fl o rH (M 00 GO CO 44 ART. 6. — Y. YENDO Literature. 1. Erhaet, G., Über subcutane Injection bei Pflanzen. (Archiv der Phar- macie, in. K. II. p. 408. 1873.) 2. Webee, f., Über die Abkürzung der Ruheperiode der Holzgewäclise durch Verletzung der Knospen, beziehungsweise Injection derselben mit AVasser. (Situngsb. d. k. k. Akad. d. Wiss. in Wien. OXX. März. 1911.) 3. Jesenko, Fr., Einige neue Verfahren die Ruheperiode der Holzgewächse abzukürzen. (Berichte d. Deut. Bot. Gesellsch. XXIX. p. 273. 1911.) 4. MiYOSHi, M., Lectures on General Botany. (Japanese edition) II. ]0- 253-4. 1905. 5. Tscheemak, f., Über die Verbreitung des Lithiums im Pflanzenreiche. (Zeitschrift f. d. landwirtschaftl. Versuchswesen in Oesterreich. p. 562. 1899. Eef. Bot. Centralb. LXXXII. p. 87. 1900.) 6 Gauneesdoefee, Joh., Das Verhalten der Pflanzen bei Vergiftungen, speciell durch Lithiumsalz. (Die landwirtschaftl. Versuchsstation, herausgegeben von Nobbe. XXXIV. p. 171. 1887. Ref. Bot. Centralb. XXXII. p. 262. 1887.) 7. Trüchot, p., Bestimmung des Lithiums mittelst des Spektroskops. (Archiv der Pharmacie, p. 561. 1874.) 8. Overton, F., Studien über die Aufnahme der Anilinfarben durch die - lebende Zelle. (Jahrb. f. Wiss. Bot. XXXIV. p. 669-701. 1900.) 9. JosT, L., Plant Physiology. (English translation p. 20 and 61. 1907.) 10. Strasbuegee, E., Botanisches Practicum. IV. Aufl. p. 113 u. 125. 1902. 11. Pfefpee, W., Pflanzenphysiologie, n. Aufl. p. 79. 194 u. 200. 1904. 12. Thoday, D., and Sykes, M. G., Preliminary Observations on the Trans- piration Current in Submerged Water-plants. (Annals of Bot. XXIII. p. 635. 1909.) 13. Snell, K., Untersuchungen über die Nahruugsaufiiahme der Wasser- pflanzen. (Flora. XCVin. p. 213-49. 1907.) iNMKcrioN' Kxri:iuMi:xT,s ox riANT.s. 45 14. Okamuka, K., On the Comparative Anatomy of AVater-plants. 1889. (Manuscript !) 15. Oltmanns, f., Morphologie imcl Biologie der Algen. 1. p. 445 u. 524, 1905. 16. ScHECHNEE, K., Zur Kenntnis des absteigenden Wasserstromes. (Anz. Kais. Akad. Wiss. Wien. XLYI. p. 272. 1909.) 17. BoEHJr, J., Umkelirmig des aufsteigenden Saftstromes. (Berichte d. Deutsch. Botan. Gesellsch. Vin. p. 311-3. 1890.) 18. WiESNER, J., Der absteigende Wasserstrom und dessen physiologische Bedeutung. (Botan. Zeitung. No. h u. 2. 1889.) 19. WiESNEE, J., über den absteigenden Transpirationsstrom. (Bot. Cen- tralb. XLni. p. 171. 1890.) 20. DixoînT, H. H. and Joly, J., The Path of the Transpiration-current. (AnQals of Bot. IX. p. 403. 1895.) 46 ART. 6. Y. YEXDO : CONTENTS. PAGE I. Introduction 1 n. Method of injection 2 m. Experiments Avith several injection solutions 3 IV. Injection experiments with several plant organs 4 Y. Injection into parenchymatous tissue 5 yi. Experiments in moist chamber G VII. Experiments on some deciduous trees in winter and early spring. 7 Vm. Experiments on some thallophytes . . 7 IX. Discussion of the results . . . . 9 1. Relation between transpiration and conduction of injected solution 9 2 Upward conduction and downward conduction 11 3. Transverse conduction 14 4. Speed of conduction of injected solutions 15 5. Conducting path of injected solutions 16 6. Injection method as a means for tracing the course of a vascular bundle 17 X. Summary 18 XI. Experimental data 20 Literature 44 Published May 15th, 1917. Y. YKNDO. INJECTION EXPERIMENTS ON PLANTS. PLATE I. Explanation of the figures. PL I. Illustrations to show the course of conduction of the injected lithium. The red shaded parts represent where lithium was detected after the injection. I.p = injection point. Fig. 1. Inflorescence of Fatsia japonica ( x — j. (Compare exp. 1). Fig. 2. Portion of a branch of Citrus nobilis with fruit [ x — ] (Compare exp. 3). Fig. 3. Leaf of Magnolia Jiypolenca ( x ). (Compare exp. 4). Fig. 4. Two leaves of Rhoeo discolor with different injection points ( x — I. \ 3 / (Compare exp. 20, 21). Fig. 5. Portion of Bicinm communis ( x — ]. (Compare exp. 6). Fig. C. Leaf of Äspiclium lacerumi x -). (Compare exp. 39). Fig. 7. Portion of Myriophyllum ( x ). The line (Z) represents the water level (Compare exp, 35). Jour. Sei. Coif., Vol. XXXVIIL. Art. 6. PI. I. T. Yendo del. Y. YKNDO. INJECTION EXPERIMENTS ON PLANTS. PLATE IL Explanation of the figures. PL n. Fig. 1. Fatsia japonica. A. Basal part of palmate veins. Through the dotted lines a, b, c, d, e the sections a', h', c', d' , e' in C were taken respectively (nat. size). B. Median longitudinal section of the same, showing the course of vascular bundles ; the coloured bandies represent those stained b}^ the injected SO4CU. The dotted lines a, h, c, d, e are like in A (diagrammatic). C. Cross sections of the same ; a', h', c', d', e' are the sections through a, b, c, d, e in A and B respectively, the coLjured bundles are like in B (diagrammatic). D. Surface view of a porti(3n of the leaf, showing tlie termination of vascular bundles being aöected by the 8O4OU injecticm (uat. size). Fig. 2. Cross section of a vascular bundle of the stem of Vtcia Faba to show xylem elements which contain aniline violet in consequence of the injection ( x 200j. Fig. o. Cross section through the rhizome of Mmianthas trifoUata to show xylem elements which contain eosin in consequence of the injection (x70). Fig. 4. Cross section of a vascular bundle of the petiole of TracJûj car pits excels I to show that much SO^Ca is contained in phloem elements in consequence of the injection ( X 200). Fig. 5. 2 stages of the unfolding bud of Cornus controversa (nat. size). In the A stage, hthium which has been injected into the stem is not detectable in the young leaves, while in the B stage it is first found in the larger leaves. (Compare exp. 56, 57.) Jour. Sei. Coll., Vol. XXXVIII., Art. 6. PI. II. Fiy. -J. Y. Yendo del. MBI. WHOI LIBRARY UH ITKY R