\ ' ' , ‘ " ' Pee tio mt , fay ‘ ‘ “ ‘) pats Hae ei i \ } ‘4 i ' i ‘ , m [Se We CC (a (CG Soe} > a bP a CME CIS MC LR al « (x «€ q@ @ Cac? @&€ Cerca Gt ¢ «eq give $c a Ck HE Ra CT @ Ce eat eg a ad ae COT EC CLE OEE CU Ge . . 3 ‘ rad «af (¢@ Sa & ( — @ueil € Calg e— ram c (< ! 4 C€< aC tee ef @ EOC CCE OC OEE CLE CEQ re CER STE COG ’ iF ecard ‘: (CUO Cae (Car ¢ nie SE ce CO AC COE OT CO « Tank « é _ Ce (CC GR aC CEC Cc | Br er. CC Gr CG. (x Cer CC Car dC CC at er (TAR AC Ca (i Ge. €C GiGm aq (Ong € aq ae yt ra eer ee & Vex fo Vb MPC ane oO C.ccame apace 4 bs) 4 pr. ig —— ) i > & “ —— ie 4 2 JR D- fF 42 A a = & = — Adel = eS 7. } el es _ ; ; AN 3 wu woe Sb == (asa - RN Retngh A : kK. wT nL eee & Sy SS . “> wis we ~ Seok. — & -o™ Sa Mt Ser eet SS i ay ~ y Se a et ee A sae a i pded Rett A dedi Ao te = ie = = = eS Se a, > art eet teties Got | | Rae > | ttd tt fhe wre 1 i wueu o°SVGee @ 7 co p72 Seey 4 rr oN < IA Ase / - NS, oN 7 ws s ~ Av a iw ey Bg _) ~~ ws La » v v ~~ Pl | ‘ Yow " — p P So y } F 4 pa Seyesss = we" v pegs OS Seni ies Be — 7 ! ~~ - E nde 7 OI. wa SAAN. GSI) AAI "ll MAA Al Cn te cy ToS yw ELM Sd ae AAI OD hah hh Ohad NASON oS oo ~ py ’ ey ce — a af i RI ONG ew > i i we, 4 , A = >» = ee w Doo fiat i fd BS Pace Bes Pins nS ear SO A US SUES eee S pte ye ~ ~ Gwe swt Uy ye? ow ww S WISeh Seen & as D AN a et at age De - ye VIN Swiss NII RSIS SIA IOI GABAA ah TD ah y ECL en a BGS RS ad Sh ~ — - = sw a AA” ) Dd daddy SY 1 gf OU te % bes! SS Sch eth “ ~= vd KS << yey | by) | wiry See Th ese J) OG we a pe : ~~ : aA Patetete db) Denth ded pth ee | SC Cee NNN OT Se. er wi A “oat het Ne ~~ = f . ww eg} No Ae WANS ~ N= we ee Nw pe SSO Uwe vee PERK eae" Shh NARA Bf Wwv° Soe ey NESSES (€ 4 a id K — S SESSSES SEIU SS Zoe Se cesy Se si eet S veer ty ise ] We ew ee New : — No Ol wy wa x 3 New we ead vad ed BS fa | “ I eS, Se ww ww ® ne pe foe you eTSY A 7 wo \ : ~) Y 4 ~~ Nat Net Nt Ba Se ad Sok I mys ead Gwe S ew) we — va Crrry ES st lt _ = Division of Fishes, U. S. National Musou 64TH CoNGRESS Dooument od Sessi HOUSE OF REPRESENTATIVES No. 2137 DEPARTMENT OF COMMERCE BULLETIN OF THE UNITED STATES BUREAU OF FISHERIES VOL. XXXVI 1917-1918 HUGH M. SMITH COMMISSIONER WASHINGTON GOVERNMENT PRINTING OFFICE Deni 4 i ee ve a ier ie te, ch 647TH ConGRESS DoouMENT od Sessi HOUSE OF REPRESENTATIVES No. 2137 DEPARTMENT OF COMMERCE BULLETIN OF THE UNITED STATES percAe OF PSHERIES VOL. XXXVI 1917-1918 HUGH M. SMITH COMMISSIONER oS \ 272826 / WASHINGTON GOVERNMENT PRINTING OFFICE 1921 IWXXK IOV oe HTIMe .MEHDUA ABIACAZAIMALAOD CONTENTS & Page THREE NEW WHITEFISHES FROM BEAR LAKE, IDAHO AND UTaH. By John Otterbein Snyder. (Mocument Sos issied Mayiy, TOLO)pisarscta eee cnlc cata eitakrow ns feidet Wot svte oe Chee eee I-10 FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES OF THE UNITED StTaTES. By Robert E. Coker. (Document/865, issued October'25;, 1919) %..s0.c0sc0eccecccscescceed ede cese veces II-9g0 LiFE HISTORY OF THE BLUE CRAB. By E. P. Churchill, Jr. (Document 870, issued November LT MLEHC Nene yarctel startet evel pencils are (oehel da.cve tee e cial nink ws Dave wisest aera sland oruseusisi ed weit Save elas APaee g1-128 SPONGES oF BEAUFORT (N. C.) HARBOR AND VICINITY. By W. C. George and H. V. Wilson. (Document’876, issued December’ 29,2919). as sciseies cee: cis tis oe nee otsiecctrostssts welsiels sive settee. « 129-180 DRAGONFLIES AND DAMSELFLIES IN RELATION TO PONDFISH CULTURE, WITH A LIST OF THOSE FOUND NEAR Fairport, lowa. By Charles Branch Wilson. (Document 882, issued August ER FES RON aN en Pheverete cisverv aye aGANeKs) sia) an ante uilekss aie nie, ayalse si dias wera leseiecainis Sidvalalurale ele Seesackie @ teacnele 181-264 BURROWING MAYFLIES OF OUR LARGER LAKES AND STREAMS. By James G. Needham. (Docu- PUEDES pISSHe Cg el yz 175 EO 20) syey-sacaet are oye Teiccavarn a) of aintese ets ol aie os 10 11 11 II 9 Ir II I II TODA A Ee eR RPO NU SIAL II Ir II 12 12 12 II 12 12 Scales'lateralline: 00s. 666. o iets ' 17 73 71 71 | 72 | 72 73 73 | 74 92 Scales above lateral lime... .......-.:-0+005s 8 8 9 9 8 | 8 8 8 8 8 Scales below lateralline.................00. 7 7 7\ 7 7 7 7 , 7 7 Scales before dorsal... 22... .6.cee eevee even 30 31 32 31 29 29 30 30 31 29 The color soon fades after death. An alcoholic specimen is brown to a point two scales above the lateral line, from where it is silvery to the midventral surface. Along the back the scales are dusky. The snout is black on the upper anterior half. The fins are without color. The intestinal canal is short and straight. There are 84 to 86 cxca just beyond the pyloris, the posterior ro or 12 extending in a single row along the intestine. The gill rakers are longest near the center of the arch, about one-half the length of the maxillary. They number 14 to 16+270r 28. Bran- 6 BULLETIN OF THE BUREAU OF FISHERIES. chiostegals 8. The air bladder is large, thin, and single lobed, extending the whole length of the visceral cavity. The peritoneum is somewhat silvery in places, but is without dark pigment. No teeth are found. Coregonus spilonotus, new species. Bonneville whitefish. Gill nets set at a depth of about a hundred feet in Bear Lake in August caught numbers of a spotted whitefish which measured from 155 to 200 millimeters in length. They were pale moss-green above, silvery on the sides, and white beneath. Spots, dusky in color, round, and somewhat larger than the pupil, extend from the occiput to the base of the caudal. These fishes differ from C. williamsoni in that the spots are smaller and more numerous, the scales are larger, and the heads longer. They were from 4 to 5 years old, and the condition of the ovaries seems to indicate that they were mature indi- viduals. At the same time and at the same depth large whitefish colored like the above, except that they were without spots, were taken on baited hooks. Besides being plain in color, these fish were much larger, 400 to 470 millimeters long; the heads were longer, the body deeper, the maxillary larger, and they were distinguished also by their general appearance. They were from 7 to 10 years old, and mature. Coregonus spilonotus. Bonneville whitefish. Locally these two forms are regarded as distinct, but a considerable series of specimens collected by Mr. Stock supplies examples intermediate in size and age, and seems to demonstrate without much doubt that they belong to the same species. The question need not be considered as settled, however, until more complete data have been obtained. This species appears to inhabit the deep water. It is to be found there as late as the month of August, and it is in the same region in January and February, when it feeds upon the eggs of other whitefish. In December, however, it migrates shoreward and spawns in shallow water. It does not enter the rivers. Type No. 83499, United States National Museum. Locality, Bear Lake, near Fish Haven, Idaho. Length 425 millimeters. Collectors, J. O. Snyder and C. L. Hubbs. Head 3.8 in length to base of caudal; depth 3.6; depth caudal peduncle 3.5 in head; snout 2.9; eye 4.8; interorbital area 3; maxillary 3.2; scales lateral series 80; between occiput and dorsal 34; above lateral line 11; below lateral line 9; dorsal rays 11; anal rr. Body deep and rather heavy, the head very large, with a long snout and broad maxillary. Gill rakers short, thick, and pointed, 6+-13 on first arch. Fins rounded; caudal small; adipose about equal in size to maxillary. Color dusky above, silvery on the sides, white below; no spots. The spots disappear with age, the head grows relatively larger, the maxillary longer, and the body deeper. The lateral series of scales numbers from 74 to 81; series above lateral line g to 11; between occiput and dorsal fin 30 to 37. The dorsal has ro to 12 rays; anal g to rr. NEW WHITEFISHES. 1£ ze re fe of 1f of of L 8 8 8 L L 8 8 or or or or or 6 It or bl og og 4 bl 14 og og 6 or or or 6 or 6 or Ir Ir or o1 11 zr II IL Szz° fz- | Siz- zee yx mi S1- 3G gi° | S6r- | Sgr- | Soz- pe z1- | Sxrr* | Scere Sgr- AY= || ox< | SLr- go- | slo: ] go- | Sgo- Bs Ere || Sor cI* gs- $5 ES\5| eS gr: | SSh-| Sob: | gb- S6r- Sr) Ores o0= ars ibs) }) Sx) (esx: go go* | t90+| go go* go* | SSo- | tgo- go- | 480-} £So-]} go- go: | LSo- | zgo- | Sgo- Lo- go- | Lo: | SLo- fz- fz- zo) fz-0 cer ost gsr zor €Sr ogr olr SLr of of L 8 6 or bh gt or or Ir or 1° be- Px Sex ®) oz* | S6r- Lat Wy @ SLi- QI: Sgo- | Sgo- er Ir zs- aa Ly- | Sov: 61° | Sgr- SS1* | Ssr- Sgo* go* go* | Sgo- go: | go- SLo- lo- go* | SLo- Szz- fz- 8 6 8 8 8 8 8 8 8 8 or II II or Ir Ir OL or or or 1g gf 6L 6L gt og SL tL og LL or or II or or or Ir Ir II Ir or IL Ir or or It Ir IL iI mI oz- Iz- | S1z- 1e- | S1z- gr: | Sgr: | Sgr- oz: | Sir- tre axe Sr Sr- S1* | Str- ys Pr*) | (Sag €1° Li: 6r- 61° gic | G6r- | Sgr: | SLr- G1: 1° gr: zr- £1* | Sor- cI- TIM Sones sxx * || were | Sro- | Str = Sz-| Sor gI- OVE SSre\Shreh Shx* | or: gi: ar go° 60° 60° go* | SLo- | Sgo- | Sgo- 60° | Sgo* | Sgo- $11* zr* zx- | Szr° zr | Srxr* | Sire | Srr- er | Srr- sss- | gS-| oS-] 65+] oS-| gS-]| Sés-| Sgs-] ses-| gs- SOP: | S6b- 1S- | Sos: os: oS* | SoS: | Sos- os: 1s° Soz- | Soz* | Soz~ | Soz- Iz: Iz* | Sie: | S6r° 1z* | Soz- or: | SLr- Lr- | Sgr: | Sgr: gi’ QI | zgr: | Sgr- | Sor- Lo+ | go* fo: | Lo: | Los | gos} Sgo*} go: | SLor | gos $So- | SSo- | SSo- go* | SSo- So- So- So- So | LSo- SLo- Lo- | SLo« | SLo+ go* | Ggo- | Sgo- | go- | Sgo- | tgo- go° go: | Slo: | SLo- | Sgo* 60- 60° | £go° 60° | ggo* tlo« | SLo- | SLo- | SLo- | SLo- | SLo- | SLo- | SLo- go° Lo- tc: fc- Sc- we Se* gz: ves gz" be | Stz- Stzo] zz+o} Sz-o] SSz+o| Sz-o| Sz-o} €Sz-o} $Sz-o] Sgz-o] SSz-o} gzo| Lz-o ge of L 8 or Ir og bl Ir or 11 1r SOP ears 41° ro Soz- | SZr- tr: [eee gr: | SSxr- £60- | Sgo- zfr: > ss+ |] oS- 1S- | Sgr: Szc- oz- Segue gr° go- | SLo* gSo- | SSo- SGo- | SLo- £go- | SLo- Lo- Lo- S£z- Sz: SLz-o| Sbz-o\- +" *""TBSIOp alojaq sapBog *- aul] [P4198] MOpEq saTeoS *- aul] [B19] BAOB saTeoS *** QUT] [8197] SITBIS verte et Seer ery sists sos Ker pesuog “-**""Tepneo qy809'T “**"]B1QU9A 44BU’T as yes0q0ad 473U8’T ""yeue IQS H “-TesIOp Ise **yeue Jo aseq 44309’T ****Tesiop JO aseq 443dUe’T “"*""Te1Q09A 0} ynoUS “*""*"TesIOp 03 NOUS “****qnd1990 03 ynouS vereesess pay yydaqr “YPM [epIqQI0s9}UT tees 9 £9 Jo ,aUIBICL “*** AreqypExeur 4}300"'T veeteeereesss anons WyyStay apounpad jepnes yidaq Peete sees deeesetesst ees Anog yydaq bite tee teeeceeeeseeeeeees oppo WQamaT zoe | gor | zie | coe | vrz | SLE | PLE | Sob | O6F | gob ect Sgt *][@UIs azIs ‘JouT;sIp AsaA sjods ‘TOMI paui 9zIs !panpqns sjodsg ‘a31v[ azIs tsjods oN ‘unm’ **"****** £poq jo yBua’y ee “*SQLONO'TIdS SAONODHAOD AO SINEUNAXNASVA 8 BULLETIN OF THE BUREAU OF FISHERIES. The gill rakers numbered 6 to 8+12 to 14. The air bladder is large and extends the entire length of the abdominal cavity. The peritoneum is immaculate. There are 135 to 140 pyloric ceca, the posterior ones extending in a single series along the digestive tract. There are 8 branchiostegals. The growth appears to be about as follows: ; Length in millimeters AV CATS OLD. 5060 cinss hese inv sh 2:n canes Sees TORT ee Cee aL eT ea cea ee 155 to 180 RATS OL oo. 5) oie) seine sim apayecesniece = oPaheresalials sieegeyera spa ieee a eve ee ae 200 ON A=) GP ADDR ORSS BESS. SUS AAS sco craccommenrT OnIgUIGg WEEMS Comes Sea ORS 255 to 260 W VEALSIONG 2 foc 0 5 s0pa> +e sp aeiagniacereyshsimdgies Sie, -,ere wich eiete tae Shots a, Sheree ae 280 to 437 Stoo Vea OIG pesos sinie.oictsie, auarn sud Pasties ae eee tae ee TE ann eee 420 to 470 Old examples are fat and weigh 2!4 pounds or more. Nothing was learned of the food of the species except that the stomachs of specimens caught during the months of January and February were stuffed with whitefish eggs. These were taken on the spawning grounds of L. gemmifer and C. abyssicola. Coregonus abyssicola, new species. Bear Lake whitefish. Small examples of this species (8 to 10 inches long) closely resemble those of C. spilonotus except that the latter are spotted. With increasing age the spots of C. spilonoius grow indistinct and finally disappear, while the maxillary and snout elongate, and the body becomes deeper. Consequently when the lack of spots fails to distinguish C. abyssicola, it may be easily separated from C. spilonotus by its much shorter maxillary. Coregonus abyssicola. Bear Lake whitefish. Local fishermen usually distinguish between spotted examples of C. spilonotus and this species, both of which they call herring, but they do not seem to suspect that the spotted fishes will grow to become the immaculate adults of C. spilonotus. Mr. Stock reports that this species is taken in sufficient numbers to ship to near-by points. It spawns from the latter part of January to early in March at a depth of about 100 feet. Examples seen alive in August were moss-green above, silvery on the sides, and white beneath. These bleached in alcohol leaving very little dark pigment, while specimens taken during the breeding season are considerably darker, indicating that they are then much more highly colored. Spawning fishes measure from 200 to 310 millimeters in length. The males are darker than the females, and the scales from the middle of the back to near the ventral surface bear mucous nodules. The females are smooth in most cases, an occasional one having small nodules on two or three rows of scales above and below the lateral line. Type No. 83500, United States National Museum. Locality, Bear Lake near Fish Haven, Id'fio. Length 310 millimeters. J. P. Stock collector. Head 4.6 in length to base of caudal; depth 4.5; depth caudal peduncle 2.8 in head; snout 3.7; eye 5.2; interorbital width 3.4; maxillary 4.1; scales lateral series 78; between occiput and dorsal 30; above lateral line 8; below lateral line 7; dorsal 10; anal 11, NEW WHITEFISHES. 9 The body is relatively slender, head short, snout short and rounded, maxillary just reaching a perpendicular through anterior margin of orbit, the latter being very angular anteriorly, and extending well forward of the iris. Gill rakers 7+11, short, thick and pointed. Fins large, the pectorals and ventrals bluntly pointed; dorsal with a straight edge; adipose much larger than maxillary; caudal deeply cleft, the lobes pointed. Color dusky above, silvery on the sides and below; no spots; scales on sides and below outlined with fine blackish dots; fins dusky, the caudal dark edged. Sex male. Each scale from the back to the level of the pectoral fin with a round, pearly mucous nodule. In a series of specimens the scales in the lateral series number from 69 to 78; between occiput and dorsal fin 25 to 30; above lateral line 8 or g; dorsal rays 10 or 11; anal g to rr. MEASUREMENTS OF TEN EXAMPLES OF COREGONUS ABYSSICOLA. Lengthiof body. o..32.ivecce es mm. . 258 234 211 235 233 245 213 223 180 206 3 3 Cy a 3 2 ie 2 g 2 Weeriptie eddie hse caciticice cle ciinieine exisivd oe «0 0. 22 0.225 | 0.22 0. 22 0-215 | 0.225] 0.23 0.215 | 0.24 ©. 205 Depth’ body) ...55....05..00% pon +24 -24 +23 +235 +24 +25 ~26 25 -25 =24 Depth caudal peduncle... . aoe 08 +075 +07 +075 +075 +074 +072 +07 +08 +07 Length caudal peduncle.................... +13 4 +145 +14 +16 +16 +16 +16 +155 Sas Length STOWE... eee eee eee eee eee «07 08 +07 +O75 +07 O75 -O7 °75 +08 -O7 Length MMARUAL Unt aceite neiesininideninn eases +05 +05 +05 +05 +052 +052 +05 +05 +055 +05 Diameter eye....... etal lateie eter win eta icra etal iat +046 +05 +048 +045 +05 +047 +052 +046 +055 +05 aterorbitalwridtha sso): hi isa keciesav ests-o sini -065 +06 +065 -065 +063 +06 -06 06 -06 -06 Depth head......... Lae ate ae aero ciate +155 -16 +15 mais +14 +155 +155 +15 -16 +14 Snout to occiput........ maot -19 «at -19 +19 +19 +19 +19 +19 +195 -18 Snout to dorsal.......... 495 st 48 +48 475 49 48 +47 +47 485 Snout to ventral........ sone 54 +55 +54 +54 +535 +56 +55 +55 +54 55 arengthi hase of dorsaly ec. ene cence eee oe -12 .125 = 105 +10 +115 +10 +115 Io 12 Io Tenigth base of anal].2 oo: a. 20s. he: - os = 095 +085 -08 +085 +10 +09 +08 085 - 085 08 Lei pitt ch ons Nepeetere aot ate eee cent eis +18 -18 165 +16 +16 -16 +155 155 +16 -16 Height anal... ni 13 12 105 = 12 12 12 11s 11s -It 105 Length pectora 205 21 19 +185 18 20 19 17 +19 19 Length ventral Ee 155 “25 135 14 +15 +15 +14 +13 135 13 SpemipthienGal es ie cia sialon cede ae ne +215 +215 +195 ~20 -20 +20 .22 195 20 2I Morsaliraysssctcs oer fee oo rete acini! Ir IL 10 ro Ir to Ir Io 1r 10 Analrays......... noe It Io 10 10 Io 10 9 Ir 10 10 Scales lateral line.......... Spar} 76 71 74 75 75 75 72 69 70 72 Scales above lateral line... ict 8 8 9 8 8 8 8 9 9 Scales below lateralline.................... 7 7 7 7 7 7 6 6 7 7 Scales beforetdlorealys .. - 2.3. cane cess 30 28 27 28 27 27 25 25 27 28 As indicated by the scales, the rate of growth appears to be as follows: - Length in millimeters. RV EATS Ol Clery tf elon eic cheese cree ercter-ketracslen ACR toon ooacee ai ble rakes say tecgees 180 if ASE NETS) 0 IN RASC CRU OES Ge BERRI DRS AeA ROR CSE Th Sena Cri otal Sa Cie 210 to 240 CSIC GNC) (7 LAS ee re AC ran 250 RV EASON Cate trees tlt eras evterrre wf, poiaie, hers eis Mtacinie ates ate meee csmeaie eas 265 MOEYCATS Gl edies favesyics stores. a yas eames ester tas Cusine mel nerve alelers st naiorsle ates atine eater 300 SRY REAUS Ol Clee sterstenranee Leccdie de myas Urn wlan he sistem oases nie erase ate arene eet 310 The gill rakers number 6 to 8+-13 to 15 on the first arch; the ceca 73 to 78; branchiostegals 7 or 8. a) ' peaplive wanted brea she Ke Bata sade ae om < Cente me dat al te “ee a sabe sive inh iy “ote nis aeIGONT I yheniy “batty 7 4 widest aotase Uy sael 402 ai vsiige it Aol Lo Wee oat tee 5% ‘ ‘ 20 2a A Re. UE Foe “e'Gr PT am oe. as woe av 47 i hye He. eS ake ; Xja- 3 ' + het 9 oat ae er, A hel 24 FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES OF THE UNITED STATES & By Robert E. Coker, Ph. D. Assistant in Charge of Scientific Inquiry, U. S. Bureau of Fisheries re jhe ae oe te anne © win ele a a te ~ é ‘ Te TT SE Sea CONTENTS: Part 1.,\Commercialfresh-watermiusdels so) ohh be ele Nols k SVAN ARTEL... Qualities of fresh-water mussel sShellgs oo. jie c.c 5 oat) uns RAMI TR Be Be siete) deed, RIB Maree AP NE os Warteties' of comimerctali smells 2 ose. .c ar) crore) chal oictora mm sa resapremtlenet tat ACO Re ae eR Ae 5 Oana CLASSE crepe ates ssafep Pep aiatce chet ecartn erases chiar sce tco) cist ate fete: ay MOURN RMR MNEND OM Tose ASR TER OU ag INMlasdeGeralts ryt} eguce monic dod: capo R UH eK e er Comp EUS OrEpLe GOOe Go tted itt cm cai BEAUTE FH EVA LOU Pareto avery nrsz0k cates eh af os 09 ores: co ve ftapy of ache ox okeahcy race’ = ARIE ME TOMEI, PTO FOG sEAOELID ere eRe a co Fials, of ele, sntialaseustacasteaiey asic sfehs Ee EES Pic enn GABOR rite ite! cic ci SRO ISRO SATU a) 6 Be b PEO OD DOD OA OUT Co ERROR Oe Sia tir EEOC COREAOE 2154.7.) Arran DUM ASE Paar CAR EOI TN tera esey oer sfokassy sich xs tucacyaleiairy sraket cha orodecahavanepayehscst oben, oaserscel cei =ian om AOI.» o «eos Ue eR EEAFYeGUL TS bel Sa Bera spe Lar eh «Poh is fot eae snt god oun shareLat eens ecoxesuegeushaieheysepee ARNE PASE OER ERE TIER 5 MSGS CT iewgedipe nao dces bee epU ee bones CubGorte AEROSOE COs no. tii a. & Rane CECE TIAGO KS STUY atefal aes fsbo 052, -sap el ou caeueyavevoiel cucisvevsuesaneivbepavain RRC QATAR PML RII: «oe 556 SHG bal al llr eh eth BER DOE ATCC Cor EOD ee Cpe C nr omrL: Sc iA toss Rao h aaa INTE Sea satis HCN PS wo ease pou, cans sue ncios = «VefoeVousachsubreuexoxeatsvoleicasyey=r=iaie acc RNRERUCL 10 MEMES MRS EPR ese os Se REET UN ET ACA ORO E Tete ags skal 024 nle eek | es0k=a fe iegnto} hehehe sYeaovan eh vs ya4s¥a/aaysas onc/y ss) 0ye 1 ROMA REECE => PERees Sty LC LEN GEVOIE Ep Joi sesc Sunde je eesioseaepe Te ceansokt fousyeseasvece ienene shakes bisvo Oa TAMU TONS AN PAERI RE co scrscavs ABS esp att tS CAR! PE OUD 3 22s joynt onesies ars oseus) ses oneseySusteraierscoiricysicraiahere cia GNA ET Lt RE OEM sieeve fake Por conameErciall SPECIES: cre ete «/oissln disse oassessyeusasucusacseuesvsus) secranvioualadeynesrtucyersicrn evel © OIA MEM Moreno, Dart 25 bresh= water MInsselsHSHEEy so c..s seu seyerssade ie tt wisirie lasCyructnr 5 ONE EMER, oRPRRE OM OSD. oR eiord ass Valterand extent of he fishery: se nooo cea rniene a ate «ein -REREIMISE TURSIS FOR TOUR Some local and temporary aspects of the fishery ............ 2 sn;e oRRTS PMOL. Bok BLAIS JO ELS Depletion of the mussel TESOURCES oo sieiciese. 2 wieueinsceye win ee ROEM E Mel PRAMS oo ess Apparatus and Methods OLASHET Yo. a aio o> wire win mie sled plan yopn\s 0iciei of RIRIENE ROTC «oe vie g seein BB Air eer dt CEO WE OOP IBOOK or ace elaine syed ae soue bs layer decks vc she 05a diy we PERE PEO II 9 oe oa ss PRATCIPLEORCAPUING yc sormieirinln eile cowie nip pe mieinbinw Inisieieicis > = SERN: SEM ESOT Description! oF Apparatus oie. cic :niovi=.m ele new leveiiehe leery eyeastare MESS AER RIE UL. « Hooksiand mode of making Pheri... ...05i cin ys cieuee meseie nine SUS y ARIEL. Bar and TMOS. 5 neu wine p nimloiose pee piclcasie SERIE AEA TS, Hep St Se LD PUSS elo een copa elie Ci ig paws ial esc Gerace ate MIS ein datas MTEL ALS. Operationvohithecrowroot bat. ius aise tin -(a enim octets polis tarts alaeeehte rete ehe Te oe as Advantages and disadvantages of the method... .)........ 0.60 c eee eee eee eee DP Met yd hae ag olla gti hea ee slone se iets uaueue obs eunaorice ta wc ee SLATS Oe SIE Onapan iat Che MELROG yee oii so jspntahe te fe eseacuoyens le sp vodedcpaie 2 chin et ahes wise «PROM Reo eed AUER JD yesrela hs) aCOval OLN) oy oF Cost CPR ea or gern aici wetioh Ge Cr Mee ORE MIC eo Deir aria cers ELA LION OLR P) Oe Umer seta a 5 caretakers roar fede st ebigh saree 5) intece eles aistaso peeve Moeae tens tis PN CVANLACESIOL LUE AEp-TEt)THEENOG «2 pra.q)teis wig |s cise sauce < += siete lelsila, emlnwesiy aleeisrein RITE Tap ANC cy any ee Ee hei Mic eRe ose. 5.6 arpa co eats wes (oe, Winle gnd evdletet sioahe mieiggenate hens ECA eae eT th i icfous, oxionsiale cheval xin creredeiclona uke kieiaieis.«, oe,aln d oreole lew a siathoeeevelete tretectoteet 6 6/7 iygrilsiriors bbrrest afore) (2 dos CR a AO ee ape AISI Ae SAEs ric, vice aCe SSR Eeg At at THE UALS) OD PISMEL YP eer atcteeioteloraeraleg/e.ciol anol lagu ce olay le ielelpipietele Mieteiyohe e// ais) n\nie is WMOLETe Ctl DINE TE a8t i PROCESSES aos eae c/a) san) a, °,0°+ leo, Sein o.s eve @iehe eteheleemieminiainle sis see > os Quadrula perplicata. Quadrula elliottii)... < Qua MemeTI 1, wore ch sleeve adda sha aes The washboard group— Washboards oy tw-asiecge > comes bsp tacit cosy MACE AeTORen sy Fey poe 8) ap h diprds shil- dees > Washboard.... .| Quadrula boykiniana. . [Bank-climber] . (Quadenia trapezoides] . 4 PCH ee cere os ceba be salcsiais cose chiles ritogonia tuberculata.............ceeceeeeeeeee XI The Lampsilis class: The mucket group— WMiicket $35 rch sun tusxs Phan peje nme cos are Fe of Lampsilis ligamentina................06 2000.0 150,00 8 APP ey [Rainbow-shell] . .| (Lampsilisiris]............ p20 0 a Re 8 Southern mucket. . .| Lampsilis ligamentina gibba. BIBLE, toreveihn avetess PEDSppUEN TS OM ecret sate a sicicle cistern orci inipisteceieineorrcistas ote Lampsilis higginsii........ oi pia ataly ala steiays 5 Lampsilis orbiculata. Wye: 9 sabac freee Fat mucket, Lake Pepin mucket.............. Lampsilis luteola.... I, XV, XVI... Southern fat mucket Lampsilis hydiana. . F Plagiola securis.......--++0+++5 [Plagiola elegarisls tsk eneecsccecasongivonessosns Tanipgilis yentrrcon@ esp as vie nd stestets a ma.ceds ae Lampsilis ovata]... ‘ Lampsilis capax].. . Lampsilis purpurata Lampsilis multiradiata].... 2.0.0.0... 000.0000 eee ‘The sand-shell group— Wellow: sard-sitell renatal trie. synieisialo cvle\ate ante Lampsilis anodontoides. .....- 2.2.00. ....000005 Slough sand-shell. . .| Lampsilis fallaciosa.... i Black sand-shell Lampsilis recta........ ; {Lampsilis subrostrata].....--- 0.0.00... cece ee ees Miscellaneous groups: The bullhead group— Bullhead Pleurobema aesopus......0+2.e.cccccceeecteeeees .| Cyprogenia irrorata......--..+.. -| Dromus dromas............+++.. .| Symphynota costata...... (Ptychobranchus phaseolus] Symphynota complanata. ...........0..00..00.. FR heel-splitter, Lampsilis alata]........ Rock pocketbook Arcidens confragosus] The elephant’s-ear group— Blephant's ears... 2. --er eens cence reece eee. Unity crasstdleris oe 5 sa ai as ciin << bo coat e Remand Let oaee Leia Mace B OMe rte occ ttle cuidate WRG BIDDOSUS Tt tite. core Saas vee ceieteres XXIL Noncommercial species— SI GALT MP hakitiion sail cielvt ores inid.c coca be Date dha Anodonta Prandishy.vieersdPuihadidis tes tdeecces Slop-bucket] ... ..| [Anodonta corpulenta]............:eeeeseeeesnes Paper-shell] .. Anodonta suborbiculata] ........... 0000. ececuee Paper-shell] .. Anodonta imbecillis]... ....-.0.sccseceeseeneeees ‘Squaw-foot].... Strophitus edentulus]......-...6. 00sec eee eee ees Spectacle-case] |... ..qsveeewrsecen nes Margaritana monodonta]... 0.6.6.6... cece eee River pearl mussel]]................. Margaritana margaritifera]...................05. xXXI Paper-shell]., 0.2.20 scecsersecoeess Teanipsilis gracilis]: cite ccs scans wsa.. auheesveuaus Paper-shell).......--..0002005 Lampsilis levissima]..........06..000ceeeee eens Paper-shell].....,...---9.--05 Pa AMLSRELOSTLS Patt |) es enw ae ep cisions anehee ee The Spheriidz] ..| (Species undetermined)... 0.0... .6 66.0 cece eevee Page. QUADRULA CLASS. This includes the niggerhead, the pimple-back, the blue-point, the washboard, and others of minor importance. Nearly all of the mussels thus classed together are short- term or summer breeders. This means that the eggs are fertilized and incubated in 20 BULLETIN OF THE BUREAU OF FISHERIES. the gill pouches of the female, passed out to become parasitic upon fish, and liberated after the period of parasitism, all within a relatively short period of weeks or months and generally during the summer season. Most of these species are tolerably restricted in their parasitism, and for this reason, as well as on account of the short breeding season, they lend themselves less readily to propagation by artificial means. The chief dependence for their conservation must now be placed upon protective measures in order to insure a plentiful supply of spawners in nature, and, as is equally important, upon efforts to promote the abundance of the fishes upon which the mussels become parasitic. These mussels were not the first to be used and appreciated, but after coming into use their popularity grew until in recent years they have constituted the greatest portion of the raw material for manufacture. Other species of mussels fell into dis- favor, but now, with the discovery of better material in the Lampsilis class, the pioneer mussels in commerce are again returning to favor. From the best to the worst there is a wide extreme, but, generally speaking, Quadrula shells are harder and of better luster and iridescence than others; these superior qualities are doubtless associated with their comparatively slow rate of growth. The individual shells show greater extremes of thickness than Lampsilis mussels (such as the mucket) so that, in cutting and finishing buttons from them, there is a relatively high proportion of waste. In addition to the relative ease of propagation, therefore, there are several practical advantages in favor of the Lampsilis mussels. NIGGERHEAD GROUP. The shells of the niggerhead group distinguish themselves from all others of the Quadrula class by combining a smooth exterior surface with a high degree of uniformity of quality. The niggerhead takes first place among the Quadrulas. The niggerhead, Quadrula ebenus (Lea) (Pls. II and III) came to be the mussel most sought, and a few years ago, at least, it was the commion standard of value. The better shells were suited to the export demand and accordingly have advanced in price. Its preeminent qualities, as compared with other species of this and the following groups, are its clear, pearly luster (equaled by only a few), the relative thickness of the iridescent portion, and its abundance in favorable streams. The nacre is of fine regular grain and lustrous white, except where iridescent. In buying mussels for button manu- facture the price is often based upon the percentage of niggerheads. The niggerhead forms 80 per cent of some mussel beds of the Mississippi but is not so common in the tributaries. It is usually restricted to the larger streams. It was once thought that the Mississippi niggerheads were better than any other, but this is not always the case. Some of the niggerheads from Arkansas are unsurpassed, especially those of the St. Francis River. Like other mussels, the niggerhead varies quite a little in form. Those which are more flat and round are preferred (PI. III, middle shell), as compared with examples which are elongate (Pl. III, upper shell) or in which there is a noticeable step-off from the thicker forward and central portions to the thinner hinder third of the shell. The shells formerly taken in the Des Moines Rapids above Keokuk were of the better character. A notable feature of the niggerhead mussel is the markedly iridescent hinder portion of the shell, and the thickness of this portion is generally better sustained in the nigger- head mussel than in any other species displaying iridescence. The buttons finished from Buu. U.S. B. F., 1917-18 PLATE IT. Upper pair: Niggerhead, Quadrula ebenus (T,ea), from Mississippi River. (See p. 20.) Lower pair: Quadrule solida (1,ea), from Mississippi River. (See p. 22.) Buy. &. 6. B. F., 1917-18. PrLats III. Upper and middle: Niggerhead, Quadrula ebenus (Lea), from Mississippi River. (See p. 20.) Lower pair: Hickory-nut, Obovaria ellipsis (1,ea), from Mississippi River. (See p. 22.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 21 the hinder portion of the shell constitute the shiny backs, as they were originally called, or iridescents, as they are now more generally termed. The iridescents are of exceptional quality and command a substantial premium in the market, where they rival the high- priced buttons made from so-called ‘‘ocean pearl,” or the shells of certain marine mollusks. While it is usually customary to finish the face of the blank which corre- sponds to the inner surface of the shell, a better product is obtained with iridescent buttons if the face corresponding to the outer surface, or back, of the shell be finished to make the face of the button. There is some difficulty not only in keeping the two sorts of blanks separate, but also in insuring that the blanks are cut entirely from the iridescent portion instead of partly from the iridescent and partly from the lustrous white surface. Since compara- tively few are obtained in any case, it is not a general practice to cut for iridescents, and most of the iridescents of commerce are cut and finished by chance, as it were, and simply sorted out in the process of grading the finished buttons. A prominent manufacturer stated at one time that a much higher price could be obtained for irides- cents if one could obtain a sufficient number upon which to build a line of trade. As it is, iridescents are generally an incidental product. As compared with shells like the Lake Pepin mucket, or the ordinary river mucket, there is considerably more waste in niggerheads, on account of the heavy hinge and teeth and the relative differences of thickness between the forward and hinder parts of the shell. For this reason particularly, relatively small niggerhead shells, from 1.5 to 2.5 inches in greatest dimension, are preferred. Such are the shells taken in the Missis- sippi about Le Claire, lowa, and in the White and St. Francis Rivers of Arkansas. In the early years of mussel fishery in any niggerhead stream, a large proportion of heavy, coarse shells were taken, and they were much less desirable. Owing to the generally depleted condition of most niggerhead beds, few large shells are now taken, but occasional specimens are found that are upward of 4 inchesin length. Some of these are of excellent quality, but there is a great deal of waste in cutting them, as most of the blanks are very thick and have to be ground to the desired degree of thinness for buttons. The relative economy in use of niggerhead mussels of different sizes is shown by the following record of tests as to number of blanks per shell and per ton: SizEs, WEIGHTS, AND BuTTON PRODUCTION FOR NIGGERHEAD SHELLS (APPROXIMATE FIGURES). Longest dimension. ne I . 1 Quantity Bs aM al fe of blanks shell. ESB: Number of mussels prea Lessthan—| Pe ton. Inches. Inches. Gross. 3 I 174; 000 II0, 000 55,000 33,000 26, 000 20, 000 15,000 10, 500 8, 500 6, 200 4,000 rye Bemus 3, 200 x RONG ARN ARN ~ ARN be MM w Pow x @ At the time of making this table only a few of the larger-sized shells were available, so that the estimates of blanks are less accurate for these sizes. 22 BULLETIN OF THE BUREAU OF FISHERIES. Because of the demand for the smaller sizes, a large number of very small shells are being marketed, many of them entirely too small to be of any service whatever. In the years just preceding 1914 there was a growing export trade in niggerhead shells of small and medium size, the price reaching $40 per ton on the river. In conse- quence of the export demand, the domestic market was diverted more and more to the inferior grade of shells. Since 1914 the domestic market has found a larger supply of the niggerheads available to it, and consequently the domestic demand for lower-grade shells has declined. During several years prior to 1914 the prices paid for niggerhead shells for domestic manufacture varied from $18 to $27 per ton; in 1919 the price per ton ranges from $40 to $80. The niggerhead mussel appears to have two spawning periods, one in spring and another in early and midsummer, but the periods are yet to be accurately defined if they are actually distinct. Like other mussels, the niggerhead is parasitic upon fish, but the only species of fish known to carry it successfully is the river herring, Pomolobus chry- sochloris. Since this fish is characteristic of deeper and swifter streams, the distribution of the niggerhead mussel is restricted accordingly. Evenin sucha large but generally sluggish river as the Illinois the niggerhead is rare, and Forbes and Richardson report that the river herring is very uncommon in that stream. There are many herring in Lake Pepin, but few niggerheads are taken there, so that other conditions must be unfavorable in this place. The niggerhead is generally found in hard, gravelly or rocky bottoms, and it is very abundant in such rapids as occur on the Mississippi. Its distribution is, how- ever, rather hard to define, since some of the larger examples have been taken in deep and slowly flowing water. The White River of Indiana, the Scioto River of Ohio, and the Duck River of Tennessee have yielded some particularly large shells. Quadrula solida (Lea) (Pl. 11) is very like Quadrula ebenus and is generally regarded by mussel fishermen as the same. The material is equal to that of the niggerhead. Although widely distributed through the Mississippi Basin, the mussel is relatively rare and small and can not be rated as of much importance. Quadrula subrotunda (1,ea) (Pl. IV) is found in the Ohio, Cumberland, and Tennessee River systems. It resembles the niggerhead, and the adults are difficult to distinguish from the latter. At Clarksville, Tenn., it is called the ‘“‘long solid,”’ and is regarded as one of the best button shells of the lower Cumberland. The hickory-nut, Obovaria ellipsis (Lea) (Pl. III) must be grouped with the nigger- head in respect to commercial qualities, although it is not closely related to it. This is perhaps the only conspicuous case in which different mussels agree closely in quality of shell while being rather distantly related in systematic characters. The hickory-nut mussel possesses a shell of essentially the same quality as the nigger- head and has sometimes been called the Missouri niggerhead. ‘The mussel differs from the Quadrulas as a class in being a long-term breeder, carrying the young in the marsupial pouches over the winter period. Closely related species are Obovaria circulus (Lea) and Obovaria retusa (Lamarck). The former is found principally in the southwestern portion of the basin. It is too small to be of much commercial significance, but will yield a few small blanks. The golf-stick, Obovaria retusa (Lamarck), is found in the Ohio, Cumberland, and Tennessee systems. Wilson and Clark report that it attains a rather large size in the Cumberland, ‘“‘but the Buu. U.S. B. F., 1917-18. PLATE IV Long solid, Quadrula subrotunda (Lea), from Cumberland River. (See p. 22.) Buu. U.S. B. F., 1917-18. PLATE V. Pimple-back, Quadrula pustulosa (ea), from Mississippi River, illustrating diversity of this species in form and sculpture. (See p. 23.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 23 deep purple of that portion of the nacre within the pallial line makes it valueless for buttons.” PIMPLE-BACK GROUP. The better shells of this group have the same general qualities as the niggerhead. ‘The best are inferior to the niggerhead only in that the backs are rough or warty, and the thickness of the tip is less well sustained. The luster is fine, and a portion of the shell is iridescent. The poorer shells of the group are worthless because of color of nacre, shape, or some other objectionable quality. Like the niggerhead, they are short-time (summer) breeders and tolerably restricted in parasitism, as far as is known. Unlike the niggerhead, they rarely, if ever, occur in such numbers as to constitute the principal species in a mussel bed. The pimple-back, Quadrula pustulosa (Lea) (Pl. V), is distributed throughout the whole Mississippi Basin in different forms and sizes but does not often attain a length greater than 2.5 inches. It occurs mixed in with other species and sometimes forms 10 per cent of the mussels in the beds. It is one of the best mussels of the Illinois River. It varies greatly as regards the size and number of pustules and rarely displays an almost entirely smooth shell. Its diversity of form is well illustrated by the several figures in Plate V. In the earlier years of the industry the pimple-back was not used. The workmen did not like it on account of the pustules on the back, which made it difficult to cut. Later, as button-making material became scarcer, it came into use and is now bought and worked up along with the niggerhead, having the same market value. The texture is firm, and the shell has a tolerably uniform thickness; since its thick- ness diminishes rather uniformly toward the tip, it can be worked up economically. It is principally used for small-sized buttons. The color of the nacre is lustrous white, and there is a fine iridescence in the hinder portion. The pimple-back spawns in early summer and midsummer, and the glochidia are parasitic, chiefly upon several species of catfish. Quadrula pustulata (Lea) (Pl. V1) is like pustulosa but is smaller and with fewer warts. It is comparatively rare and is not distinguished commercially from pustulosa. Quadrula cooperiana (1,ea) (Pl. V1) is a more southern form of pimple-back found in the Cumberland and Tennessee systems. It is called pimple-back, but, unlike the north- ern form, the nacre may be white or from a pale to a deep shade of pink. A blank of from 30 to 36 lines can be cut from the white shells. The maple-leaf, Quadrula lachrymosa (Lea) (Pl. VII), is not found in great quan- tities but occurs in small numbers among other mussels; for this reason it was once known as the “stranger.” The material is of a good, white luster and firm texture, but, owing to the thin tips, about half of the blanks can be used only as ‘‘tips,” which is the commercial term for blanks less than one-twentieth of an inch in thickness.? A small proportion of iridescents is obtained, and, but for the thinnish tips and knobby back, the shell would be equal to that of the niggerhead. When found in considerable numbers in the shell piles at the cutting plants they are sometimes sorted out and cut separately. @ See footnote, p. 17. 24 BULLETIN OF THE BUREAU OF FISHERIES. The shells are of about the same size as the niggerhead. Those 2.5 to 3 inches long are fairly large. Mr. Boepple reported that the Scioto and Duck Rivers yielded examples of maple-leaf 4 to 5 inches long. The maple-leaf mussels probably spawn in early summer and midsummer. Quadrula fragosa (Conrad) (Pl. VII) is a very rare species of the maple-leaf, more quadrate in form than Quadrula lachrymosa; while Tritogonia nobilis (Conrad), having a similar external appearance, is confused with the maple-leaf. The monkey-face, Quadrula metanevra (Rafinesque) (PI. VII), is found in relatively small numbers. It occurs infrequently in the large mussel beds, but is usually found near the bank or outside the main beds. Owing to the large pustules and the very uneven outer surface it is difficult to cut, but with careful handling it may be cut into a few blanks of small size which are of excellent quality. The shell is often used to advantage for cutting one 24 or 30 line button from each shell. In value the shell is sometimes classed with pig-toes. "The spawning time is early or midsummer. The rabbit’s foot, Quadrula cylindrica (Say) (Pl. VII), is a very long and narrow form that is familiar to the fishermen of the southern portion of the Mississippi Basin. It is too narrow, convex, and uneven of surface to be of value for button manufacture. One or the other of the purple pimple-backs, Quadrula granifera (Lea) (Pl. VI) and Quadrula tuberculata (Rafinesque) (PI. VI), is found in most large rivers of the Mississippi and Great Lakes Basins, but they are not generally distinguished. The species tuberculata is flattish and is probably found more often in the smaller or tribu- tary streams. Both species are found in small numbers scattered among others. Owing to the purple color of the nacre, the shells have no commercial value. The layers are said to split apart in cutting. The shells have a very attractive appear- ance and will take a beautiful polish when finished as souvenirs. ‘The mussels are of value in the rivers, since they produce a relatively high number of pearls. Scarcely a tuberculata could be taken in the Grand River in Michigan in 1909 without finding some sort of pearl formation. The spawning period is early summer. The three-horned warty-back, Obliquaria reflexa (Rafinesque) (PI. VI), is not at all closely related to the pimple-backs or purple warty-backs. It has one row of large knobs on each shell, and the knobs are remarkable in that those of the two sides are not opposite, but alternate in position; the species can not, therefore, be mistaken for any other. The three-horned warty-back is found in small quantities along with other mussels. The forward portion of the shell is thick, the tip thin. The form and the knobs are objectionable, and the size is not large, but the texture is good, and the nacre is clear and white and makes first-grade button material. The species is widely distributed in the Mississippi drainage and elsewhere. It appears to be a summer breeder. PIG-TOE GROUP. This group is rather limited both in variety of species and, except in certain streams, in general abundance of the mussels: None of the mussels is of the best quality. The pig-toe, Quadrula undata (Barnes) (Pl. VIII), is found in small quantities, prin- cipally in the Mississippi, and also in some of its tributaries. White River, Ind., has examples of unusually large sizes. While the material is somewhat similar to that PLATE VI. a, Pimple-back, Quadrula pustulata (I.ea) (p. 23); b and c, purple pimple-back, Quadrula granifera (Lea), from Mississippi River (p. 24); d, three-horned warty-back, Obliquaria reflexa (Rafinesque) (p. 24); e, _pimple-back, Quadrula cooperiana (1,ea), from Cumberland River (p. 23); f, purple pimple-back, Quadrula tuberculata (Rafinesque) (p. 24). BULL. Uo. Bebe monz— 1S Pate VII. a, Monkey-face, Quadrula melanevra (Rafinesque) (p. 24); 6, maple-leaf, Quadrula lachrymosa (Lea) (p. 23); ¢, maple-leaf, Quad- rula fragosa (Conrad), from Ohio River (p. 24); d, pig-toe, Quadrula plena (1,ea), from Cumberland River (p. 25); e, rabbit’s foot, Quadrula cylindrica (Say), from White River, Ark, (p, 24). Buu. U.S. B. F., 1917-18. PLATE VIII. Upper pair: Ohio River pig-toe, Quadrula obliqua (Lamarck), from Ohio River. (See p. 25.) Lower pair: Pig-toe, Quadrula undata (Barnes), from Illinois River. (See p. 24.) Buu. U.S. B. F., 1917-18. TehyNaNeY JOSS Upper pair: Flat niggerhead, Quadrula coccinea (Conrad), from Fox River. (See Dp. 25.) At bottom: Wabash pig-toe, Quadrula ruhoinosa (1.ea), from Cedar Creek, Ind. (See p. 25.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 25 of Q. ebenus, its commercial value is not of first rank; the shell can not be worked up as economically as that of some other species and is usually cut into small-sized buttons. The pig-toe (formerly called Quadrula trigona) is a summer breeder. Ohio River pig-toe, Quadrula obliqua (Lamarck) (Pl. VIII), is the commonest mussel in the Ohio River. It is found throughout its entire length as far as mussel beds extend. It often forms 80 per cent of the mussels in a bed. In some places in the Cumberland River this is the most common mussel, and the shells from the Cumberland River are superior to those of the Ohio. It is found also in the Illinois River, but is rarely, if ever, seen in the Mississippi. The commercial value of its shell is not high. When first marketed from the Ohio River it brought only $1 to $2 per ton, and buyers were difficult to obtain at that price. In 1910 the price reached $8 per ton, but it has been higher since ($12 to $13 in 1914, and about $30 in 1919). ‘The material has a poor luster and is chalky, and only the butt part of the shell can be used, as the iridescent portion of the shell is too thin. The nacre is often marked with green spots, and many shells are eroded at the umbones, these qualities being more evident in the shells from the upper portion of the river. Quadrula plena (Lea) (Pl. VII) and Quadrula pyramidata (Lea) (Pl. XII) are two species that are distinguished taxonomically, but they can not be differentiated com- mercially from the other pig-toes. The flat niggerhead, Quadrula coccinea (Conrad) (Pl. IX) and the Wabash pig-toe, Quadrula rubiginosa (Lea) (Pl. IX), are found principally in the small rivers. Both species are reported from the Grand River, Mich., and coccinea is common in the James River, S. Dak. They are not uncommon in the small outlets of lakes of the northern States. While the extreme forms are readily recognized, the two species run into each other (in external appearance),so that they are often confused. Rubiginosa has a pro- nounced posterior ridge. In shape the shells are sometimes rather circular (as is especially true of coccinea) or rhomboidal. They are more compressed than the ordinary run of niggerheads or pig-toes and have a relatively light-colored epidermis. The shells are somewhat puzzling to fishermen but are often called flat niggerheads or thin nigger- heads. They are not, however, closely related to the niggerheads in scientific characters or commercial qualities. Mr. Southall states that the nacre is sometimes rather soft, like the pig-toe, and sometimes very hard, like the bullhead (Pleurobema esopus). It so happens that coccinea (but not rubiginosa) is placed by Dr. Ortmann in the genus Pleurobema. It is probable, therefore, that the commercial qualities, as well as the scientific positions of the two species, are quite distinct; but, in the lack of final information and with the present confusion in common parlance, they are mentioned together. The nacre of both species is of fair quality, but the iridescent part is too thin to be used. Coccinea often has a pink nacre, and in this case it is sometimes called “pink niggerhead.”’ The spawning times of both species are probably early summer and midsummer. BLUE-POINT GROUP. We come now to species that are much larger in size, always with rough, ridged backs and with a quality of nacre somewhat inferior to the shells of the niggerhead. Many of them are well esteemed, especially because so many blanks can be cut from the single shell, and because they are adapted for the larger sizes of buttons. 26 BULLETIN OF THE BUREAU OF FISHERIES. The species are not so restricted in parasitism as the niggerhead and the pimple- backs, and plicata, at least, is carried by several of the game fishes. The blue-point, Quadrula plicata (Say) (Pl. X), and the three-ridge, Quadrula wndu- lata (Barnes) (Pl. X), two very similar mussels, called by the mussel fishermen blue-point and three-ridge, are among the most widely distributed species in the whole Mississippi Basin, being found in most of the rivers and larger creeks in different sizes and forms. Plicata is the thicker species, with heavier umbones, common in the deeper and more sluggish waters, while undulata is flattish and characteristic of headwater or tributary streams. ‘The species seem to intergrade, so that it is frequently difficult or impossible to distinguish them. The clammers do not seem to recognize the two species, but apply the term three-ridge or blue-point indiscriminately. Generally speaking, these mussels, even when clear of spots, work with a good deal of waste, on account of the heavy hinge and teeth, and they yield a considerable number of second and third grade buttons, although some buttons of very good quality are also produced, including a few iridescents. Blue-points, three-ridges, and washboards (see below) were worth about $12 per ton in 1914, and about $30 in 1919. The commercial value of the shells varies greatly in different rivers and creeks. In the Mississippi River, for example, the young mussels can be sold with the nigger- heads. ‘The value of the shell decreases as the mussel grows older. The shell loses iridescence and becomes more brittle and hard, and consequently difficult to work up; the layers lose their firmness of attachment, so that they split off easily. Old shells, moreover, are frequently spotted. It is found in manufacture that the iridescence of tips from these shells is enhanced by the processes of bleaching. These mussels spawn in midsummer. In the streams of the gulf drainage in Florida, Georgia, and westward these species are replaced by Quadrula perplicata (Conrad), Quadrula elliotti (Lea), and Quadrula neislerii (Lea). Quadrula perplicata occurs in the Cumberland under the common name of round-lake shell. WASHBOARD GROUP. This group comprises, practically speaking, only a single species. The washboard, Quadrula heros (Say) (Pl. XI), is the largest and heaviest species of mussel in the Mississippi Basin. One example 8 inches wide and 11 inches long and weighing over 4 pounds with the flesh, was collected by the late J. F Boepple in the Salt River, Ky. The empty shell weighed about 314 pounds. The washboard is found chiefly in large rivers in quiet, deep water and on gravel and mud bottoms. ‘The Wabash and Illinois Rivers have the highest percentages of washboard, although there are beds in the Ohio River where this species forms nearly 50 per cent of the catch. The shell is generally of an oval outline and more or less elongated. It is valued chiefly because of its large size, making it suitable for cutting the largest-sized buttons. The material is tough and the grain uniform and regular. The iridescent part breaks easily in sawing, owing to the undulations on the back. The nacre is usually discolored with yellowish or greenish spots, and the older the shell the larger are the spots. We have received pink shells from the Illinois River. Young mussels from 3 to 4 inches long have only a few spots or none, and the iridescent part is as thick as in the older mussels, being thick enough for buttons. This part of the washboard is very similar to the Burrs on baie) DOl7—Loe PLATE X. Upper pair: Three-ridge, Quadrula undulata (Barnes), from James River, S. Dak. (See p. 26.) Lower pair: Blue-point, Quadrule plicata (Say), from Illinois River. (See p. 26.) Bui. U.S. B. F., 1917-18. PLATE XI. Upper and middle: Washboard, Quadrula heros (Say), from Mississippi River. (See p. 26.) At bottom: Buckhorn, Trilogonia tuberculata (Barnes), from Mississippi River. (See p. 27-) Buy. U.S. B. F., 1917-18. PLATE XII. Upper and middle: Bank-climber, Quadrula trapezoides (1ea), from White River, Ark. (See p. 27.) At bottom: Pig-toe, Quadrula pyramidata (Lea), from Cumberland River. (See p. 25.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 27 mother-of-pearl of the ocean. The washboard spawns in the late summer and early fall. In places in Florida, Georgia, and Alabama heros is replaced by Quadrula boykiniana (Lea). The bank-climber, Quadrula trapezoides (Lea) (Pl. XII), should be mentioned in this connection. It is found in streams flowing into the Gulf of Mexico, from Alabama to Texas, and northward in the Mississippi system to Tennessee and Arkansas. The shell has a deep purple nacre and is quite valueless for manufacture. Possibly it yields a proportion of pearls. It isavery familiar shell in Louisiana, Arkansas, and eastern Texas. The buckhorn, Tritogonia tuberculata (Barnes) (Pl. XI), is, perhaps, better named in connection with the washboard than anywhere else. It has a naturally white nacre of good texture and quality, but is often spotted. It is thinnish at the tip and has a very rough back; some shells have a pinkish tinge. It has also been called pistol-grip, a name appropriate to the form of elongate examples. There is a short form character- istic of males and a much more elongate form common to females. It is found widely in the Mississippi and Gulf drainages and is reported as a summer breeder. LAMPSILIS CLASS. Such familiar and valuable shells as the mucket, the Lake Pepin mucket, and the sand shells are representatives of this class of mussels. In many respects they are quite distinct from the Quadrulas. In commercial quality there is a wider range, not only between the species composing the class, but even within the individual species in most cases. The highest-priced shells of all are of a Lampsilis species, while some of the most worthless paper-shells are species of the same genus. Muckets may possess excellent qualities, or again they may be pink or otherwise inferior; some pocketbooks are good, some are worthless. Fat muckets from one region may sell for scarcely less than niggerheads, while those from another locality would not be looked upon with the thought of marketing. The species of Quadrula, asa rule, have more uniformity wherever found; some are better than others, but when a Quadrula is found there is a reasonable presumption that it is a shell of a certain grade, according to its species. The primary commercial difference between Quadrula and Lampsilis is that the latter rarely shows any marked iridescence. Sometimes iridescent qualities are referred to, but this generally means merely an unusually bright luster. On the other hand, Lampsilis mussels have a more uniform thickness, and therefore yield a larger number of blanks per ton than any of the Quadrulas. Some of these mussels are not surpassed in texture and luster, as will appear, and therefore this class of shells has been growing in favor in recent years. As previously mentioned, the raw materials first used in fresh-water button manufacture were species of Lampsilis. The Lampsilis mussels are more rapid of growth than the Quadrulus, and they are long-term breeders. In the latter part of the summer, as a rule, the marsupial pouches are filled with eggs which develop into glochidia, and in this condition all or a large proportion of them are held over the winter. Glochidia can be found in the gills of the females at almost any season of the year. July and August, principally, constitute the 28 BULLETIN OF THE BUREAU OF FISHERIES. intermediate season, when few glochidia can be found. These months, it is important to note, are the height of the breeding season for most of the Quadrulas. MUCKET GROUP. We include in this group three abundant and important shells, the mucket, the fat mucket, and the southern mucket, and three species less abundant and less readily dis- tinguished. The mucket, Lampsilis ligamentina (Lamarck) (Pls. I and XIII), is one of the most generally distributed mussels in the United States. The commercial value of the shells varies primarily with the rivers from which they are taken. The muckets of the Mississippi River have not been highly esteemed. The butt (or heavy) portion is considered too chalky, and the tips are rather thin. In many of the shells the nacre is pink in color, which greatly reduces the value. ‘There are, however, some places in the river where the quality is superior. The muckets of the Wabash River have been considered very fine, but they are now rather scarce. The Yellow River, Ind., has produced excellent muckets. In the Ohio River better muckets are found higher up the river. Mr. Boepple reported them abundant and of good quality at Marietta, Ohio. They are abundant in the Green River, Ky., and are of excellent quality in the Little Barren River, Ky. Mr. Boepple also stated that he had found muckets which approached marine shell in luster in the Cottonwood River, Kans. Muckets are comparatively scarce in the Illinois River, but some of the fishermen believe they are becoming more numerous. The mucket can be found in almost any sort of stream, and the best shells are usually found where there is a good current, but this is not a universal rule. In the Grand River, Mich., where the current is good, the muckets have an excellent luster, but too large a proportion of the shell is very thin. The material works up well; it is soft and has a straight grain, although in old shells the nacre splits, and the nearer to the hinge one is cutting the worse this trouble becomes. Some muckets have excellent luster and clear color, but these qualities vary with the locality. The color varies even in the same bed; pink muckets and white muckets are found side by side, and the cause of this difference in color is as yet unex- plained. The values on the basis of ton price in 1914 and 1919, respectively, may be stated as approximately $17 and $45. The mucket may, perhaps, liberate its glochidia to some extent in the fall, but principally in spring and early summer. It has a relatively wide range of fish hosts, principally among the game fishes. The little rainbow-shell, Lampsilis iris (Lea) (Pl. XIII), with its bright-green, broken rays on a yellowish shell, is often mistaken for a young mucket. It is found in the Ohio River system, and also in the streams of Illinois, Wisconsin, Michigan, and eastward. The southern mucket or yellow-back mucket, Lampsilis ligamentina gibba Simpson (Pl. XIV), one of the finest of all shells, differs from the common mucket in being shorter and more compressed. The shell is therefore flatter and the thickness more even; the texture and luster are unsurpassed, and the material works easily and econom- ically. This form is found in streams south of the Ohio River and perhaps, too, in that BuLL. U.S. B. F., 1917=18. PLATE XITL. Upper and middle: Mucket, Lampsilis ligamentina (Iamarck). (See p. 28.) At bottom: Rainbow-shell, Lampsilis iris (Lea), from Tippecanoe River, Ind. (See p Bui. U. §. B. F., 1917-18. PLATE XIV. Upper pair: Southern mucket, Lampsilis ligamentina gibba Simpson, from Black River, Ark. (See p. 28.) Middle pair: Higgin’s eye, Lampsilis higoinsii (Lea), from Mississippi River. (See DP. 29.) At bottom: Lampsilis orbiculata (Hildreth), from Cumberland River. (See p. 29.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 29 river. Our collections have been principally from the Cumberland, Clinch, and Holston Rivers. The Higgin’s eye, Lampsilis higginsti (Lea) (Pl. XIV) @ is a rather uncommon species, but a few may be found in almost any carload of mixed shells from the Ohio or Illinois Rivers, or from the middle section of the Mississippi River. The nacre is often yellowish or pinkish, but white shells are of first grade, with good thickness and luster. Lampsilis orbiculata (Hildreth) (Pl. XIV) very closely resembles Lampsilis higginsit, but is more southern in its distribution. Its qualities are about the same as those of the preceding species. The fat mucket or Lake Pepin mucket, Lampsilis luteola (Lamarck) (Pls. I, KV and XVI), would not have been classed some years ago as an important commercial shell. Now, since the Lake Pepin form has come into use, the species is considered one of the best. It is widely distributed over the upper half of the Mississippi Basin, the Great Lakes drainage, and, according to Simpson, the entire Dominion of Canada east of the Rocky Mountains. Its occurrence as an economic form is quite restricted. The fat mucket is rarely found in rapidly moving water but is adapted to slow or still water. It is the principal shell of the lakes of the Middle West, both in the Mississippi and the Great Lakes drainage, but such waters are not generally suited to the production of commercial shell. The fat muckets of the lakes are usually some- what dwarfed and inflated, with the shell strongly curved, and too thin to produce more than one or two small blanks, if any. In some cases they might be confused with dwarfed pocketbooks. The species is also found in rivers and creeks, but usually close along shore, perhaps well up on the banks and out of the main current. Examples from such locations generally have an inflated but more elongate form, the forward part (or, in commercial terms, the butt) of the shell being thick enough for blanks, while from one-third to one-half of the shell, or more, is so thin as to produce only tips at best. In this form the fat mucket is sometimes much like the slough sand-shell in superficial appearance. Such shells can be used, but they are not always valued highly. Within recent years the beds of Lakes Pepin and St. Croix have been discovered commercially. These lakes yield a very distinct type of fat mucket, which is remarkably even in thickness, with a surface relatively flat in males, and even in females much less curved than usual. Practically the entire surface can be cut into blanks which are of a suitable thickness. No other shell of any species cuts with so little waste, either as to the proportionate weight of shell that is thrown away after cutting out the blanks or as to the small amount to be ground from the blanks in reducing them to a proper thickness for buttons. The shell is clear white, the texture good, and the luster leaves nothing to be desired. The Lake Pepin mucket brought in 1914 and 1919, respec- tively, about $20 and $35 per ton. In the brilliancy and the extent of the iridescent portion the Lake Pepin mucket is not quite equal to the niggerhead and pimple-back, but a measure of iridescence is found, and in pearly character of the nacre it is fully equal to any other. As regards economy, it has been found that 14 or 15 pounds of blanks are avery good return from 100 pounds of niggerhead shells, while more than 20 pounds of blanks may be obtained 9 This species and the following may be more closely related to the pocketbook than to the mucket, as Ortmann holds for orbiculata. Our classification is only for shell qualities, of course. 110307°—21——3 30 BULLETIN OF THE BUREAU OF FISHERIES. from 100 pounds of Lake Pepin muckets. This shell is, therefore, the nearest approach to an ideal button shell now found among fresh-water mussels. The rate of growth is relatively rapid. At the Fairport station mussels of this species have grown to a length of more than 1 inch within six months after the date of infection upon the fish. These were in floating crates in the river. The age of com- mercial shells can not yet be positively stated but it is probably from 4 to 6 years.% Like most others of the genus, it is a long-term breeder. Its fish hosts are the common game fishes, such as the basses, crappies, stnfishes, perches, and sand pikes. Some have been grown in ponds at the Fairport station to a length of about 1 inch in a season, and very thin buttons have been cut from such shells at the end of the second growing season. The Lake Pepin mucket lends itself to methods of artificial propagation better than any other species. Lampsilis hydiana (Lea) may be called the southern fat mucket, being found in the lower portion of the Mississippi Basin. The specimens we have had might easily be confused with the Lake Pepin form and appear to resemble it in qualities of shell. It occurs in Louisiana, Texas, Arkansas, and neighboring States. The butterfly, Plagiola securis (Lea) (Pl. XV), presents another case of a mussel which must be placed far from its systematic position. Its shell qualities place it more nearly with the muckets than with any others. It is a well-known mussel of the larger streams of the Mississippi and Ohio drainages and is reported from Alabama. Its beautiful form and markings give it the name of butterfly. Mr. Boepple remarked that comparatively few females were found, and that they are of much lower commercial value than the males, on account of being so much thicker and lacking in luster. The shells of the males are very flat and have a white color and good luster, with a rather uniform thickness over most of the shell. There are few places where 100 pounds can be obtained, but they are often met with in mixed shipments, and are valued. It is a long-term breeder, and its most common fish host is the fresh-water drum, or sheepshead, A plodinotus grunniens. A smaller species, the deer-toe, Plagiola elegans (Lea), is very common, but rarely attains a size sufficient for cutting blanks. POCKETBOOK GROUP. The pocketbook, Lampsilis ventricosa (Barnes) (Pls. XV and XVI), is a large and very inflated mussel found throughout the Mississippi and Great Lakes drainages (as well as in the Nelson River), in large and small rivers, and in some lakes. It is one of the species most familiar to the fishermen and most readily obtained. It is found in gravel bars or sandy bottom, sometimes alongshore and sometimes in the deeper water. The commercial value of the pocketbook is generally rather low. The shells of the male are better than those of the female. There are some rivers in which the pocket- book becomes a very good shell for button manufacture. Mr. Boepple had an example from the Yellow River of Indiana, from one side of which 52 18-line blanks were obtained, all of which would make good buttons. The Yellow River specimens are among the best, since the hinder or tip portion of the shell is thick enough for buttons. As a rule better shells are found in small rivers and creeks than in the large rivers. a At ages of 314 and 4}4 years some Lake Pepin muckets reared in a pond at the Fairport station yielded ra to 22 14-line blanks over 234 lines in thickness. Buu. U.S. B. F., 1917-18. PLATE XV. Upper left: Lake Pepin mucket, Lam/psilis ludteola (Lamarck), male. (See p. 29.) Upper right: Lake Pepin mucket, Lam psilis luteola (Lamarck ), female. (See p. 2y.) Lower left: Butterfly, Plagiola securis (ea). (See P. 30.) Lower right: Pocketbook, Lampsilis ventricosa (Barnes). (See Pp. 30.) Buy. U.S. B. F., 1917-18. IRATE Revell Upper: Lake Pepin mucket, Lampsilis luteola (Lamarck). (See p. 29.) Middle: Yellow sand-shell, Lampsilis anodontoides (Iea). (See p. 31.) Lower: Pocketbook, Lam/psilis ventricosa (Barnes). (See p. 30.) Bui. U.S. B. F., 1917-18. PLATE XVII. Upper: Grandma, Lampsilis ovata (Say), from Cumberland River. (See p. 31.) Lower Pockethook, Lampsilis capax (Green), from Mississippi River. (See p. 31.) Bury. U. S. B: B:, 1917-18: PLATE XVIII Upper pair: Purply, Lambsilis purpurata (Lamarck), from Kiamichi River, Okla. (See p. At bottom: Lampsilis multiradiata (1,ea), from Stone River, Tenn. (See p. 3t-) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 31 Often the forward or butt part of the shell is somewhat chalky and comprises scarcely more than one-fourth of the shell surface, while the remaining three-fourths is too thin and brittle. A noteworthy feature of the shell is the lateral hinge, which has a beautiful pearly luster. The cardinal teeth, too, have an attractive form and are used in the manufacture of cheap jewelry. Lampsilis ovata (Say) (Pl. XVII), the southern pocketbook or grandma, is not ordinarily distinguished from ventricosa. It is found in the Ohio River and tributaries, as in the Clinch, Holston, and Cumberland Rivers. It is rather thinner and inferior to the common northern pocketbook. Lampstlis (Proptera) capax (Green) (Pl. XVII), also called pocketbook and confused with the others, is not closely related to them in spite of its superficial resemblance. It is quite too thin for value in button manufacture. The purply Lampsilis (Proptera) purpurata (Lamarck) (Pl. XVIII) is probably related to capax. It is very familiar to shellers in Texas, Louisiana and Arkansas, but the thinness of the shell, as well as the deep purple nacre, makes it unfit for the trade. It is one of the most beautiful shells. Another species which looks something like a young pocketbook, but which never attains so large a size, is Lampsilis multiradiata (Lea) (Pl. XVIII) of the Ohio drainage and southern Michigan. Its shining greenish-yellow, bright-rayed shell is very attrac- tive to the eye. SAND-SHELL GROUP. There are three sand-shells, the yellow, the slough, and the black. The yellow sand-shell, Lampsilis anodontoides (Lea) (Pls. I, XVI and XIX), is the most highly prized of all commercial shells. It is never very abundant, but it is probably the most widely distributed of all the species discussed. Its distribution as given by Simpson is; “‘ Entire Mississippi drainage, except, probably, the upper Missouri. All the Gulf drainage from the Withlacoochee River, Fla., to the Rio Grande, and into Mexico.” The common name is derived from the clear yellow or brownish-yellow exterior color. These shells are too valuable for use in button manufacture; consequently they are always sorted out. Many tons of them are bought from the fishermen on the rivers to be used for export. Even those that reach the factories in mixed cars are sorted out in the yard to be sold again to shell jobbers. In very recent years, however, due to war conditions, many sand-shells have been cut into buttons in domestic manufacture. Some years ago it was said that the sand-shells were shipped chiefly to France; in the few years preceding 1914 the greater part seemed to have been destined for the German market, and the price on the rivers in 1913 was $60 per ton. The export was neces- sarily interrupted in 1914, but in the early part of 1911 the writer was informed of an offer of $92.50 per ton f. o. b. New York, and a consular report from Hamburg quoted these shells at prices equivalent to from $108 to $151 per long ton, when niggerheads were quoted in the same market at $54, and muckets at from $47 to $119. About three pairs of sand-shells (the shells from three mussels) usually make a pound, so that the mussels were worth at least 1 cent apiece on the river and, at the date of the consular report, 2 cents apiece or more in Hamburg. In 1919, with some export demand, yellow sand-shells bring about $90 per ton. : The shell owes its value to the following characteristics: (1) Its luster and pearly qualities are almost if not quite equal to the marine shells; (2) its texture is smooth and 32 BULLETIN OF THE BUREAU OF FISHERIES. firm; (3) the shape of the shell is long and straight, so that pieces suitable for knife han- dies or other novelty objects can easily be cut from it; and (4) the comparative uni- formity of thickness and the light hinge make the shell yield the best return in proportion to its weight. This species is found in small quantity mixed with other mussels in the principal mussel beds or on the more sandy or gravelly shoals. It seems also to live well in muddy bottoms. Like others of the genus, it is a long-term breeder, but is, so far as known, very restricted in parasitism. No other hosts than the several species of gars seem to carry it well, but there is reason to believe that the large-mouth black bass may also serve as host. It is a peculiar fact that the two species of most restricted parasitism are the niggerhead and the yellow sand-shell. We know only one host for the niggerhead, yet it is a very abundant mussel; there are several species of gars, and they are quite plentiful; but the sand-shell is never very numerous in any locality. A noticeable characteristic of the yellow sand-shells is the habit of wandering about on the bottom; for they travel more than the mussels of any other species. The yellow sand-shells are frequently observed to crawl up on the shoals with the rising water, and it is common report that after the subsidence of floods they may sometimes be found far out in the swamps. The sand-shell has a relatively rapid rate of growth, probably attaining a market size in four to six years. Its growth is undoubtedly more rapid in the South, as in Arkan- sas, than in the North. If any species should prove adapted for commercial rearing in private waters, it would seem that the yellow sand-shell and the Lake Pepin mucket offer the best promise. The slough sand-shell, Lampsilis jallaciosa (Smith) Simpson (Pl. XIX), is similar to the yellow sand-shell but is generally smaller and rather too thin. Its geographic range is wide, and it is said to have been much more abundant formerly than now. ‘There are few places in which the slough sand-shell is at all numerous. It is common in the lower part of the Illinois River, and is a very familiar shell in the Wabash River, where it is mistakenly called bank-climber. The black sand-shell, Lampsilis recta (Lamarck) (Pl. XIX), is also widely distributed in the Mississippi River and the Alabama River drainages, the Red River of the North, and the St. Lawrence system. It is found principally in the upper part of the Missis- sippi Basin. The shell is generally more compressed and heavier than the other sand-shells. The nacre has an excellent surface, but its economic qualities are variable. Often the nacre is deeply colored, pink, salmon, or purplish. White shells are the rule in the Mississippi and in some other streams, and many of them are of very superior quality. They were sometimes exported with the yellow sand-shell. The black sand-shell has a peculiarly good luster and pearliness and even iridescence; some of the most beautiful specimens the author has seen were, however, condemned by manufacturers as too brittle and as otherwise inferior. Lampsilis subrostrata (Say) (Pl. XTX) might sometimes be confused with the black sand-shell or more easily, perhaps, with very dark-colored slough sand-shells. It is entirely too thin to be of value. Bury. U.S. B. F., 1917-18. PLATE XIX. Yellow sand-shell, Lampsilis anodontoides (Lea), from Miss: Slough sand-shell, Lampsilis fallaciosa (Smith) Simpson, from Mississipy Black sand-shell, Lampsilis recta (Lamarck), from M ississippi River. , Lampsilis subrostrata (Say), from Mississippi River. ssippi River. (See p. 3 River. Pp. 32. (See p. 32.) ee P. 32.) 15}.0)0 0 ORS YID SYN) clean Kop ey aos ho PLATE XX. Top: Bullhead, Pleuwrobema aesopus (Green), from Wabash River. (See p. 33- Middle pair: Dromedary mussel, Dromus dromas (Lea), from Cumberland River. (See p. 33.) ower pair; Fan-shell, Cyprogenia irrorata (Lea). from Cumberland River. (See p. 33.) Buy. U.S. B. F., 1917-18. PLATE XXI. Upper pair: Kidney-shell, Pivchobranchus phaseolus (Hildreth), from Obey River, Ky. (See p. 33.) Lower pair: Rock pocketbook, Arcidens confragosus (Say), {rom Mississippi River. (See p. 34.) Bui. U.S. B. F., 1917-18 PLATE XXII. Upper left and middle right; Pink heel-splitter, Lampsilis alata (Say), from Mississippi River. (See p. 33.) Lower left and lower right: White heel-splitter, Symphynota comblanata (Barnes), from Kankakee River. (See p. 34 -) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 33 MISCELLANEOUS GROUPS. There remain for brief mention about a dozen species which have little in common. None is very numerous, generally speaking, but all are more or less familiar to fishermen and manufacturers, while nearly all enter to some extent into button manufacture, although they are rarely bought intentionally. Some, at least, have certain good qualities. BULLHEAD GROUP. The bullhead, Pleurobema esopus (Green) (Pl. XX), is a yellow-back, thick, knobby shell, with nacre of excellent whiteness and luster. It is, however, exceedingly hard in texture, difficult, therefore, to cut, and injurious to the saws. After the blank is cut, the button is finished readily and takes a good polish. The fishermen in some locali- ties have facetiously dubbed this shell the “clear profit,’ because it can be thrown in to add weight at a profit to themselves and, supposedly, at a loss to the buyer. Sheep- nose is another name, referring to its form. It is a mussel of rather wide distribution in the Ohio drainage and eastern part of the Mississippi Basin northward. The fan-shell, or ringed warty-back, Cyprogenia irrorata (Lea) (Pl. XX), isa smaller and flatter mussel of good form and appearance, but its qualities are reported to be sim- ilar to those of the bullhead. It occurs in the Ohio drainage, especially Tennessee and Kentucky. The dromedary mussel, Dromus dromas (Lea) (Pl. XX), is somewhat like the fan-shell, but without the low protuberances on the back of that shell. It is found in the Tennessee and Cumberland River systems. In appearance it suggests the southern mucket, but its qualities, so far as known, correspond to those of the fan-shell. The kidney-shell, Ptychobranchus phaseolus (Hildreth) (Pl. XXI), is a much more elongate shell, with hard nacre and an undesirable steely luster. It is found in the Ohio drainage and is reported to extend northward to Michigan and southwest- ward to Louisiana, Arkansas, and Kansas. HEEL-SPLITTER GROUP. The white heel-splitter, or pancake, Symphynota complanata (Barnes) (Pl. XXII), is of wide distribution in the upper Mississippi and Ohio drainages, the upper St. Law- rence drainage, and the Mackenzie River. It has a large, fine surface, but unfortunately, the shell is nearly always thin. In some localities it becomes very large and of suitable thickness but is brittle. The buttons can be finished with good luster, but the shell is liable to split into pieces when the blanks are being cut. It is said that they can be cut readily when fresh from the river and before the shell has so dried out as to be checking and splitting. In some places the name elephant’s ear is applied to this species. The name is appropriate enough, except that it has already been so generally applied to another species to be discussed later. The fluted shell, Symphynota costata (Rafinesque), a comparatively thin-shelled mussel of wide distribution, has recently come into use from certain streams in Wisconsin. 34 BULLETIN OF THE BUREAU OF FISHERIES. The pink heel-splitter, Lampsiis alata (Say) (Pl. XXII), is mentioned in this connection only because of the confusion of names. It has about the shape and thick- — ness of the white heel-splitter but is always purple or pink and is worthless for button manufacture. The beauty of the nacre and of the teeth makes it useful in novelty work. It occurs in the Mississippi drainage at least as far south as Arkansas, as well as in the drainages of the St. Lawrence River and the Red River of the North. The rock-pocketbook, or bastard shell, Arcidens confragosus (Say) (Pl. X XI), has little resemblance to the white heel-splitter in form, but its nacre seems to be of the same character. It has the rough exterior of a blue-point, with the inflation of a pocket- book, which accounts for itscommon names. It is probably related to the Symphynotas. The species is rare but widely distributed. ELEPHANT’S-EAR GROUP. This is the last group of commercial mussels, and the shells possess peculiar features, good and bad. There are only two species to be considered, the elephant’s ear, Unio crassidens (Lamarck) (Pl. XXIII), and the spike or lady-finger, Unio gibbosus (Barnes) (Pl. XXIII). The former is distributed through the Mississippi drainage generally, and occurs also in the Alabama, Tombigbee, and Chattahoochee Rivers. The spike has a very similar distribution, but extends into the St. Lawrence and its tributaries, being common, for example, in the small streams of Michigan. As regards economic qualities, the characterization of the two species must be the same, except that the spike has generally a poorer form and is more often inferior in texture. The elephant’s ear is broader, more rectangular in form, and heavier. The spike, as the name implies, is more elongate and thinner at the tip. The entire shell is sometimes very thin, as found in small streams. Both species generally have an appearance described as ‘‘solid,” with a thick anterior (butt) portion and often with a very uneven surface. The color may be deep purple, reddish, or salmon, or occasionally white, and is often particularly beautiful. Unfortunately, coloration of any kind detracts from the commercial quality of a shell. The thickness is good, but the most favorable feature of the shell is its texture, which is probably equal to that of marine shells. A manufacturer stated that cutting elephant’s ears as compared with ordinary hard, white shells was like sawing a cake of firm soap as compared with sawing a board, and quoted a trade maxim: ‘When you find a pink shell you find a good shell’’; that is, a shell which cuts well, although its color may make it undesirable. Analysis shows that the color is a feature of the organic matter in the shell and not of the crystalline or lime content. No clue has yet been obtained as to the nature of the coloring matter, nor has any entirely satisfactory method of bleaching been discovered, unless quite recently. If the color could be removed from the nacre by a cheap process that would not injure the texture or luster of the shell, the elephant’s ear would become a most popular material. Elephant’s ears, when purchased in mixed lots, can be used to advantage for the production of smoked-pearl buttons, if stained with silver nitrate. Occasionally car- load lots of this shell have been purchased on the rivers toward the close of the season, but the practice has been discouraged by the tendency of shellers to throw in all manner of pink, purple, and otherwise useless material. Both species are used in making novelties. These mussels are probably short-term summer breeders. Buty. U.S. B. F., 1917-18. PLATE XXIII. Upper pair: Elephant’s ear, Unio crassidens (Lamarck), from Mississippi River. (See p. 34.) Lower pair: Spike, Unio gibbosus (Barnes), from Mississippi River. (See p. 34.) JPioree, WhSh 18h Lee, Weems. PLATE XOXITV. Upper and middle: Floater, Anodonta grandis (Say), from James River, S. Dak. (See p. 35.) Lower: River pearl mussel, Margaritana margaritifera (Linnaeus), from Freeport, Me. (See p. 35) Bui. U.S. B. F.,-1917-18. PLATE XXV. Upper pair: Squaw-foot, Strophitus edentulus (Say). (See p. 35.) Lower pair: Paper-shell, Anodonta imbecillis (Say). (See p. 35.) BULL. U.S. B. F., 1917-18. PLATE XXVI. Upper: Paper-shell, Anodonta suborbiculata (Say). (See p. .35) A dis gytee Lower pair: Spectacle-case, Margaritana monodonta (Say), from Mississippi River. (See p. .35) The three small shells in upper left-hand corner of plate are examples of Spheride, not fresh-water mussels. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 35 NONCOMMERCIAL SPECIES. We have completed the list of mussels of some economic value, numbering about 41 in all. There are approximately 533 species of the Unionidae in North America (Simpson). Therefore, though granting a few omissions in the present list for species from localities little studied, it is discovered that the great majority of the American species are not of economic use. It is desirable to mention only a few of those forms which are most abundant and most familiar to shellers, or of particular interest for other reasons. Most of the noncommercial species are not numerous, and some are quite rare and restricted in distribution. Even in the economic investigations of fresh-water mussels it is necessary to devote some attention to the useless forms, because the study of such mussels often throws significant light upon the propagation, habits, and other relations of the economic forms. The floaters, Anodonta grandis (Say) (Pl. XXIV) and other species, and the slop- bucket, Anodonta corpulenta (Cooper), are familiar shells of the slack waters along river shores or in sloughs. The shell is very thin and brittle and entirely useless. In one of the overflow lakes near Fairport there are found numbers of the very large and com- pressed Anodonta suborbiculata (Say) (Pl. XXVI). The most interesting species of Anodonta is the small, elongate, and paper-shelled Anodonta imbecillis (Say) (Pl. XXV), which has been found to develop without parasitism. Another species which has been observed to develop without parasitism, but which has also been found to develop with a stage of parasitism on fish, is the squaw-foot, Strophitus edentulus (Say) (Pl. XXV), which is widely distributed in the Mississippi and Atlantic drainages. "The Anodontas are not only thin-shelled but also are marked by the entire absence of hinge teeth; Strophitus has no teeth but a thickened hinge line. The shell is thicker than that of Anodontas but is too thin for commercial use; the nacre is generally of a dirty-yellowish color. The spectacle-case is a thin-shelled mussel of the elongated form suggested by its name. ‘The species is Margaritana monodonta (Say) (Pl. XXVI), of the Ohio, Cum- berland, and Tennessee systems, but rare in the Illinois River. Its interest consists in the fact that it is the nearest relative of the river pearl mussel, Margaritana margar- itifera (Linnzus) (Pl. XXIV), the principal pearl-bearing mussel of Europe and New England. The paper-shells, as the name indicates, are too thin for utility, although of attrac- tive appearance. They are very common and of wide distribution through the Missis- sippi Basin and elsewhere. The principal species are Lampsilis gracilis (Barnes) and Lampsilis (Proptera) levissima (Lea). Just as there are minnows among the fishes which are very small, even when adults, so there are small forms of mussels; an instance is afforded by the rainbow-shell, men- tioned on an earlier page. One of the smallest of all true fresh-water mussels is Lampsilis parva (Barnes), which grows abundantly in the ponds at the Fairport station. It is narrow in form and never attains a length much exceeding 1 inch. Much of the skepticism of practical men regarding the results of investigations of the life history of mussels has been due to a lack of acquaintance with the small species. Time and again shellers or others have found reason to believe that mussels bring forth their young in an advanced stage of development, because they have found supposedly young niggerheads attached to, or inclosed within the dead shells of old niggerheads. 36 BULLETIN OF THE BUREAU OF FISHERIES. In nearly every case the little mollusks thus found are not the young of larger mussels, nor are they mussels of any character. It should be clearly and generally understood that, while all the larger clams of the rivers are mussels, there are clams which are not mussels at all, being markedly different in structure and development. There is no common name for these, but they pertain to several species of Spherium, Musculium, and Corneocyclas (Pisidium) (Pl. XXVI). The usual diameter is 0.25 to 0.5 inch. They are frequently found attached to stones or to old shells in the mud. A dead shell may be found to be full of them. The Unionide (fresh-water mussels) and the Spheeriide live side by side in our rivers, but each family is more nearly related to different families of ocean shells than to each other. SUMMARY. Of the North American species of fresh-water mussels, more than 500 in all, we have named 66 as more or less familiar to the fishermen, but of these only 41 can be classed as having commercial value in the shell trade. Some of the others are valuable as producers of pearls. Looking at the 41 used in manufacture more closely we find only 17 that are of real importance at the present time. It is desirable to name these separately. Quadrula class: Lampsilis class: Niggerhead. Mucket. Hickory-nut. Southern mucket. Pimple-back. Lake Pepin mucket (fat mucket). Maple-leaf. Butterfly. Monkey-face. Pocketbook. Pig-toe. Yellow sand-shell. Ohio River pig-toe. Black sand-shell. Blue-point. Three-ridge. Washboard. Many manufacturers or buyers would reduce this list by omitting several of the species, but there would probably be little agreement as to the species to be eliminated. The best of all species at the present time are the yellow sand-shell, the nigger- head, the southern mucket, and the Lake Pepin mucket, the last three being of approx- imately equal value. The yellow sand-shell has been used entirely for export and commands a price nearly double that of other species. Many niggerheads were exported a few years ago, causing a distinctly advanced price. There is a great deal of variation in quality among the several species. Some shells are better for one purpose, while others are better for another. Within the species there is variation according to the locality in which the mussels have grown. A mussel may have a shell of good quality in one stream and of poor quality in another. The differences may apply to color, luster, texture (firm, chalky, brittle, or hard), and form (shape and thickness). Within the same mussel bed there may be differences in quality in the same species. We may find side by side pink muckets, white muckets, etc. The shells of the Quadrula class show more uniformity in quality over the entire region of distribution than those of the Lampsilis class. Iridescence is best shown in the niggerhead and pimpleback groups, but only a very small percentage of truly iridescent buttons can be obtained in any case. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 37 The requisite qualities are pearliness, luster, absence of color, and a texture firm but not too hard; and these qualities are as good in the better Lampsilis shells as in the better Quadrulas. The most common defects for nearly all shell species are spotting or staining, due, in many cases, undoubtedly, to parasites, and natural nacre colors, such as pink, salmon, or purple. Stains are most common in sluggish rivers. Colored shells seem more prevalent in clear, shallow streams, but no universal rule has been observed. Many of the tiny shells found attached to old shells are not young fresh-water mussels, as is often assumed by shellers, but belong to an entirely different family of bivalves. PUBLICATIONS TREATING MUSSEL RESOURCES OF VARIOUS STREAMS. Coker, Ropert E. 1912. Mussel resources of the Holston and Clinch Rivers of eastern Tennessee. U.S. Bureau of Fisheries Document No. 765, 13 p. Washington. 1915. Mussel resources of the Tensas River of Louisiana. Economic Circular No. 14, U. S. Bureau of Fisheries, 7 p. Washington. Coxer, Rosert E., and SouTHaLt, Joun B. 1915. Mussel resources in tributaries of the upper Missouri River. (With description of shell found in the James River, Huron, $. Dak., July 27, 1913.) Appendix IV, Report, U. S. Com- missioner of Fisheries for 1914, 17 p., 1 pl., 1 map. Washington. DANGLADE, ERNEST. 1914. The Mussel Resources of the Illinois River. Appendix VI, Report, U. S. Commissioner of Fisheries for 1913, 48 p., 6 pl., including 1 map. Washington. ELDRIDGE, JOHN A. 1914. The mussel fishery of the Fox River. Appendix VII, 8 p., Ibid. IsELy, F. B. 1914. Mussel streams of eastern Oklahoma. Economic Circular No. 9, U. S. Bureau of Fisheries, 6p. Washington. MEEK, S. E., and CrarK, H. Warton. 1912. The mussels of the Big Buffalo Fork of White River, Arkansas. U.S. Bureau of Fisheries Document No. 759, 20 p. Washington. Sura, Austin F. 1913. The mussel fisheries of Caddo Lake and the Cypress and Sulphur Rivers of Texas and Louisi- ana. Economic Circular No. 6, U. S. Bureau of Fisheries, 10 p. Washington. [Urrersack, W. I.] 1914. Mussel resources in Missouri. Economic Circular No. ro, U. S. Bureau of Fisheries, 6 p. Washington. Witson, CHARLES B., and CLarK, H. Watron. 1912. Mussel beds of the Cumberland River in rg1z. Economic Circular No. 1, U. S. Bureau of Fisheries, 4 p. Washington. 1912. The mussel fauna of the Maumee River. U. S. Bureau of Fisheries Document No. 757; 72p.,2pl. Washington. 1912. The mussel fauna of the Kankakee Basin. U. S. Bureau of Fisheries Document No. 758 52 p-,1pl., r chart. Washington. 1914. The mussels of the Cumberland River and its tributaries. U.S. Bureau of Fisheries Docu- ment No. 781, 63 p., r pl. Washington. Witson, CHaries B., and DANGLADE, ERNEST. 1912. Mussels of central and northern Minnesota. Economic Circular No. 3, U. S. Bureau of Fish- eries, 6p. Washington. 1914. The mussel fauna of central and northern Minnesota. Appendix V Report, U. S. Commis- sioner of Fisheries for 1913, 26 p., 1 map. Washington. Part 2. FRESH-WATER MUSSEL FISHERY.* VALUE AND EXTENT OF THE FISHERY. The fresh-water mussel fishery is older than the fresh-water pearl-button industry, since the mussels have been taken in the search for pearls since 1857 at least, although but locally and irregularly. ‘The importance of the fishery, in its two phases of pearling and shelling, dates from the beginning of the manufacturing industry, in 1891. It is interesting to note that at the present time the value of the pearl product is equal to about one-half that of the shell product. In some streams, chiefly the smaller ones, the pearls bring a better return to the fishermen than the shells, the Black River of Arkansas being a notable instance; but generally the value of the shells is considerably greater than the return from the pearls; the usual ratio is about 2 to 1. The present paper is intended to refer primarily to the shelling industry and to give somewhat briefly an account of the territory and methods of the fishery. Since the pearls are usually taken incidentally in preparing the shells, the pearling methods are essentially the same, except that in regions where pearling is almost the exclusive object the practice of cooking out is not followed, owing to the belief that heat is detrimental to the pearls. In the shell fishery many noneconomic mussels are taken and cooked out along with the commercial shells with the hope that additional pearls may be found. It would be of interest to compare the shell production of earlier years with the more recent statistical data for the mussel fishery. The earliest available estimates of the mussel fishery are contained in Statistics of the Fisheries of the Interior Waters of the United States, by Hugh M. Smith. The quantity of mussel shells taken in 1894 is stated at 195,500 pounds (equivalent to 97.75 tons), having a value of $2,737. The small quantity of shells and the high unit value indicate that the industry was in a very rudimentary condition then, when few shells were required, and those bought were by the pound. It is well known from other sources that, owing to the great abundance of shells in proportion to the market demands, the price soon reached a low level, about $5 per ton, fluctuating from $4 to $10; but the supply was such that the fishermen made better wages then than at the present time, when the price received per ton is many times higher. Smith © states that in 1897, 3,502 tons of shells were taken in Iowa and Illinois with a value of $40,408, and in 1898, 3,641 tons with a value of $37,008. Almost the entire fishery was within the limits of these two States at that time. It would appear that the average price per ton was about $11.50 in 1897 and about $1o in 1898. A census report for 1889 shows that 23,824 tons of shells were taken, with a value of $216,404 (average price, $9.04 per ton), and a census report for 1908 gives the tonnage as 38,133, worth $386,000 ($10.02 per ton). @ Ernest Danglade, formerly assistant in the Bureau of Fisheries, aided materially in the preparation of the description of the methods of fishery. > U. S. Commission of Fish and Fisheries: Report of the Commissioner for the year ending June 30, 1896. Washington, 1897. ¢ Smith, Hugh M.: The mussel fishery and pearl-button industry of the Mississippi River. Bulletin, U.S. Fish Commis- sion for 1898, Vol. XVIII, p. 289-314. Washington, 1899. 38 FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 39 The latest data are found in detailed statistical reports of the Bureau of Fisheries covering the territory of mussel fishery by sections.? In 1912 about one-third of the territory produced 19,715 tons, valued at $294,606, and another third produced in the following year 23,317 tons, at a value of $382,210. The remaining third, including the Mississippi River, canvassed for 1914, a year of very poor fishery, yielded 8,539 tons, bringing to the fishermen $148,960. It seems a fair estimate that the total production of shells varies in different years from 40,000 to 60,000 tons, with a value of from $800,000 to upward of $1,000,000 (not including the value of pearls sold). The shells consumed in manufacture in 1912 aggregated 55,671 tons.? The territory surveyed for 1912 comprised the southern portion of the Gulf drainage and the southern portion of the Mississippi Basin up to and including the Ohio River and up to, but not including, the Missouri or its tributaries. Arkansas was credited with nearly one-half of the total production. The average price of shells in Arkansas was nearly $20 per ton, but the average ton price for the entire region was $14.94 (1912). The territory covered in 1913 was that north of the Ohio and east of the Mississippi. The Illinois and Rock Rivers were credited with more than half of the total product for this territory and year, and, although the average price of shells in those streams was about $14 per ton, the average ton price for the territory and year (1913) was $16.82. For 1914 the survey covered the Mississippi River and its tributaries from Kansas northward. The Mississippi River produced more than three-fourths of the tonnage, and the average price per ton for that river was $19.47 per ton, as against $17.44 average ton price for all streams covered in the survey for that year. Combining the three surveys, it is found that the average price per ton of all shells was $16 for that period. Shells have advanced so materially in price during the years from 1914 to 1919 that the average price per ton is now about 100 per cent higher. The history of the fresh-water mussel fishery since the beginning of the button industry has been marked by a continual extension of territory from the point of origin on the Mississippi River near Muscatine, lowa. The rate of spread has been directly correlated with the rate of depletion of the more central territory. At various times the Wabash, the Ohio, the Illinois, and the rivers of Arkansas (White, Black, and St. Francis) have taken turn as the principal seats of mussel fishery. That the spread of the fishery has not been to the eastward and southward alone is shown by the fact that the fisheries have been prosecuted in South Dakota and Kansas and extensively in Minnesota. Unquestionably Lake Pepin in the Mississippi River between Minnesota and Wisconsin has recently been yielding a greater quantity of shells per linear mile than any other stream or portion of a stream. The principal mussel streams are listed in the table following. The total value of the pearl and shell product is shown, as well as the year of the survey, and the last column indicates what proportion of the total income of the mussel fishery in each stream is derived from pearls. The data are taken from the statistical bulletins previously cited. a [Fresh-water mussel fishery of streams tributary to the Gulf of Mexicofrom the Ohio River southward in 1912.) Report, U. S. Commissioner of Fisheries for 1914, p. 26-30. Washington. [Fresh-water mussel fishery of streams tributary to the Great Lakes and the Ohio and Mississippi Rivers northof the Ohio and east of the Mississippi River in 1913.) Report, U. S. Commissioner of Fisheries for 1915, p. 64-69. Washington. [Fresh-water mussel fishery of the Mississippi River and its western tributaries from Kansas northward in 1914.) Report, U.S. Commissioner of Fisheries for 1916, p. 55-57. Washington. b [Fresh-water pearl-button industry of the United States in 1912.) Report, U. S. Commissioner of Fisheries for 1914, p. 31-34. Washington. 40 BULLETIN OF THE BUREAU OF FISHERIES. Total value of B shell and J pearl Pearl River. State. Year. | products | products to in total. fishermen at the Tiver. Per cent. 1914 $176, 510 29 I9I3 150, 696 2I 1913 128, 692 3r 1912 122,748 38 I912 118, 891 10 eeerstehe ee 1912 68,726 65 abash........ 1913 611991 35 Bast Fork.................+.....-.---{ Indiama............... 1913 45) 0! 19 StiGroix-..... 1913 37)032 63 1912 29, 769 17 Riss ibies ubleinig'e en dial oa%p Slate eine Mamionhyere chia. | MERON Dee Els eee ae ae 1913 23,970 25 Cimmbernd ss. ict iiss ccccecce ce cevdewacess| Genmessee: IENUCKY = ® hy y A > 4 7) .. . “0 . a. % : iY alt Py iP p Ted 7 - Ne ol ae Ls +. Kel ~7 FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 41 SOME LOCAL AND TEMPORARY ASPECTS OF THE FISHERY. The mussel fishery is a permanent and important industry, and in scarcely any locality where a shell fishery was once established has it ever been entirely abandoned; yet the intensity of the fishery in any locality may vary from year to year, as it is sub- ject to a variety of influences, such as the demand for shells of certain qualities, the stages of the river, the condition of local industries or of the button industry, and the degree of exhaustion of the material. The natural movement of the territory of fishery from regions more or less depleted to virgin streams has been discussed on page 39. A limitation upon the rate of expansion is imposed by the cost of transportation of the product from its original source to points of manufacture; but the principal determining factor in this respect is the quality and abundance of the material, since cutting plants will follow the fishery if justified by the nature of the available material. There may be given, first, some illustrations of the extent of the fishery in certain localities as they have come more or less directly under the observation of the writer. The statistical bulletins published by the Bureau show the value of the product from the several streams, each taken as a whole, but the high productivity of certain beds before the process of depletion has advanced is not brought out in such reports. In the Illinois River in 1909 there were estimated to be about 2,600 boats engaged in the mussel fishery. One hundred or more boats would be engaged upon the same bed at one time, and, as a consequence, cutting plants sprang up all along the river. In 1911 the writer counted 125 mussel outfits lying idle upon the banks at one point, Merdosia, III., while some 15 or 20 were engaged upon the river at varying distances from the town and mostly out of view. The exhaustive effect of the fishery of the three years preceding 1911 accounted in some measure for the number of idle boats. In 1912 only about 400 boats could find occupation in the mussel fishery on this river, although a more effective method of capture (the dip net) had been devised. In each of the two years 1912 and 1913, according to reliable information, about $20,000 worth of shells and pearls were marketed at Madison, Ark., a village of about 300 inhabitants. This represented the product of beds largely depleted but yielding shells of high quality. In 1910, 1,600 tons of shells, principally niggerheads, were taken in the rapids above Keokuk, Iowa, mostly within a stretch of 4 or 5 miles, and these shells represented a value of about $30,000. In 1912 a little over 400 tons were taken in the same region. This bed has been lost to the industry since the submergence of the rapids following the construction of the dam. In the immediate region of Ie Claire, lowa, and Port Byron, Ill., in 1910, about 700 tons of shells were taken, representing a value of some $14,000. Fishing at Le Claire, lowa, began about 1897, and the large catch of 1910 was due to the condition of very low water, enabling the fishermen while wading to pick by hand the shells which could not be taken in ordinary stages because of the rocky character of the bottom. These are not insignificant figures, considering that the harvest was reaped without expense of planting or cultivating. Such harvests can not, however, be often repeated, since the rate of removal exceeds the rate of natural replenishment. 42 BULLETIN OF THE BUREAU OF FISHERIES. The variable quality of the several species of mussels is discussed on page 16 of Part I in connection with the commercial qualities of shell. Only a word is necessary in this connection in regard to the geographic aspect of the subject and its effect upon the fishery. The profitableness of shelling in any locality is determined largely by the quality of the more abundant species. Some species are nearly always rare, or, at least, are never the dominant species of mussel beds. Nevertheless it is a striking feature of the mussel fishery that in different streams or in different portions of the same stream different species may dominate. We may have niggerhead streams, mucket streams, pig-toe streams, etc., but we do not expect to find maple-leaf streams, buckhorn streams, or butterfly streams, although these forms are widely distributed. In that portion of the Mississippi constituting Lake Pepin, the Lake Pepin mucket comprises as much as 60 per cent of the catch from many beds, in spite of the large number of other common species present. Since this shell as found in Lake Pepin is of the best quality in so many respects, the region of the lake may be expected to be the scene of active fishery as long as the beds are reasonably productive. Certain portions of the Mississippi River may show from 50 to 85 per cent of niggerheads against all other species combined. At Le Claire, Pleasant Valley, and other points above Davenport, Iowa, 75 or 80 per cent of niggerheads are reported. Counts of shell piles above Keokuk, October, 1912, showed 80 per cent niggerheads, and 10 per cent monkey-faces, while 7 other species constituted the remaining 10 per cent. There were a few discards not included in the count. A similar predominance of niggerheads is observed in Arkansas, especially in the White and St. Francis Rivers. In the St. Francis near Madison, Ark., in 1913, about 16 species were taken, but 75 per cent were niggerheads. In such regions one may expect a steady fishery until the beds are nearly exhausted. It is reported that in the lower Pearl River of Mississippi and Louisi- ana the niggerheads constitute more than 99 per cent of the mussels in beds that have been fished, but the quality of the shell has not been definitely ascertained; if the report of percentage is correct, these beds are the most remarkable known for the predomi- nartice of one species. In the Illinois River Danglade ¢ found that in various beds blue-points may con- stitute 50 to 60 per cent of the catch, washboards 23 to 50 per cent, and warty-backs as much as 31 percent. His observations regarding the river as a whole are succinctly expressed by the statement: “The Illinois is distinctly a washboard, blue-point, and warty-back river.”’ In piles of shells taken near Havana, Ill., washboards and blue-points constituted 95 per cent of the shells, while nine other species constituted the remaining 5 per cent. The fortune of the fishery in this stream will necessarily fluctuate with the demand for that class of shells, which is not at all constant. At times there may be a strong demand for blue-points and washboards for the making of buttons of the larger sizes; but, except with such a demand or with an excessive call for the higher-class shells, the market will not be the most favorable. In the Ohio River, the Ohio River pig-toe may constitute as much as 80 per cent of the mussels of a bed, but this species is never in high favor; mixed in with the pre- dominant species, there is always a certain number of mussels of other species, some of which may be of superior quality. @ Danglade, Emest: The mussel resources of the Illinois River. Appendix VI, Report, U. S. Commissioner of Fisheries for 1913, 48 pp., 6 pl. including 1 map. Washington, ror14. FRESH-WATER MUSSELS AND MUSSEL, INDUSTRIES. 43 In the Cumberland River * the Ohio River pig-toe is reported to run as high as 95 per cent of the catch from one bed, while the valuable southern mucket is found in the proportion of 40 per cent in some beds in this river. An unusual export demand, causing a high price for the best shells, will tend to throw the domestic trade back upon second-grade material, and thus stimulate the fishery in regions of inferior-shell product. On the other hand, a slack in exports such as now prevails makes the best shells more readily available to American manufac- turers and discourages the fishery for poorer shells. Taking the principal mussel streams previously listed, we find the average price per ton for shells on the bank in the years 1912-1914 to be as follows. These figures are of value now only as indicating the relative values of shells from the several rivers. RELATIVE VALUES OF SHELLS IN VARIOUS STREAMS, BASED ON SURVEYS OF 1912-1914. a Average : Average River. Year. ton price. River. Year. ton price. ROM so win ciare tatets le cle Sunflower! River, Miss). 0 =. 0- vecmep uneh veces 3-50 2.56 5,500 Butterhy.. AIS. F<. Jorn | White Rivet) Ark.) i 0. 20h cee. odes 3-00 2.44 9,000 Lake Pepin mucket.. ....| Mississippi River, Lake City, Minn... 3-17 1.92 10,000 Maple-leaf........... ....| White River, Ark 7 DA BNL 2-54 2.00 II, 000 Monkey-face........ ....| Mississippi River, Fairport, Iowa... . ee 3-00 2. 43 8,000 Mucket...... Airey see BENE Sree aecsoace adie aa 4:67 2.80 6, 800 Do.. = : 4-47 2. 60 5,000 Do.. .| Wapsie River, Waverly, Iowa . 4-98 2-95 5,500 Niggerhead .| Mississippi River, Fairport, Iowa ae 3-56 a. 81 4,000 DOU. cise ....{ Sunflower River, Miss ...............005 ad 2-95 2.20 6,500 BOGS cat as ....| White River, Ark....... as a 2.58 2.14 9,000 Pig-toe....... B do apr 2.23 2.00 13,000 Pimple-back. : 2.10 2.00 16,000 Pocketbook. . 7 Sa 490 3°35 5,000 TOG acpeo ....| Wapsie River, Waverly, Mayas ce stieaedsccho nae 5-10 3-29 4) 000 Three-ridge............ sia A] OM RRA VER ) WWAS ceil ct clots aniardia ite Hee Bee 4-33 3-33 ‘Ay 800 wegbout ak Ari Mississippi River, Fairport, Iowa....... a5 5.75 412 2,000 Wale eeaitme sites ....| Mississippi River, Grafton, Ill........... is! 4: 40 3-10 4) 500 Yallow sand-shell. . ....| Mississippi River, Fairport, Iowa....... oa 4°72 2-33 6,000 POs e eeadeitassnene panealar cise ers acs GO iss atacctuving eiteins stolonle stepisis's onto s atalino orate = eis 4-23 2.00 9,000 Many of the shellers are nomadic and therefore move readily with their launches from a region of poor fishery to a better locality. It is often the case, however, that in times of low water, when the mussels are easily obtained, the farm hands, miscellaneous laborers, and others engage temporarily in shelling, using any kind of available equipment or collecting by hand. It is in such cases that good beds are often rapidly and seriously depleted. A noteworthy difficulty encountered in some places where the quality of shells is good is the high cost of transportation. In regions remote from manufacturing centers it is therefore advisable to have cutting plants, so that the expense of shipping the waste portions of the shell may be obviated. A cutting machine costs about $16. The cost of a plant of 12 machines, complete, with all equipment except power, was about $400, as computed in 1914. In 1919 a cutting machine of improved type sells for $28 anda 12-machine outfit is estimated at $725.% DEPLETION OF THE MUSSEL RESOURCES. It has been mentioned that the extension of the fishery has been directed by the depletion of the mussel beds in the regions first worked. Generally speaking, it may be said that no stream or region has been entirely exhausted, but wherever a mussel fishery has once existed it has continued in operation tg this time, although in reduced activity and with much irregularity. The history of a shell bed in typical cases may be described. When first fished, there is usually a large proportion of very old shells which are coarse and heavy and often much eroded. After the first year or two the yield of the bed is chiefly the medium- @ In referring to values and costs the writer feels obliged throughout this paper to give figures, when available, for both the years 1914 and 1919. While it can not be assumed that prices of 1919 are normal, it seems reasonably sure that prices will not return to the level of 1914. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 45 sized shells of the best quality. In the last stages the larger shells constitute a gradually decreasing proportion, while the smallest shells, the very infants, indeed, form a steadily increasing proportion. Here is a fatal defect of the present manner of fishery. The rate of depletion is automatically accelerated, since the fishermen are taking two or three for one. Where formerly from 5,000 to 10,000 mussels, more or less, constituted a ton, at a Jater time, when small shells prevail, each ton may represent some 30,000 mussels, as has been determined by repeated counts made by the writer in localities where the small shells are being marketed. The fact that the sheller can now take but a few hun- dred pounds a day, as against the former catches of half a ton or more, might lead one upon first thought to suppose that the beds thus automatically receive a measure of protection. The very contrary, however, may in practice be the case. ‘The continued rise in price stimulates the sheller to save everything, and the last stages of impoverished fishery are thus the more exhaustive. When shelling in a depleted locality becomes quite unremunerative, it may be practically entirely abandoned and almost forgotten, until after some years, it is found that the growth and natural reproduction of the mussels have so replenished the bed that it has again become a profitable one, and general shelling is resumed. Usually, however, the local shellers keep engaged, though irregularly, upon the same bed, or more exhaustive methods are employed, and the cases of natural recuperation are therefore conspicuous by their rarity. In view of the conditions just described, the Bureau has advocated the compulsory closing of portions of rivers for periods of years, in order that the mussel beds might have such a condition of rest and freedom from all injurious disturbances that the process . of replenishment would be assured. It has also urged the adoption of size-limit regula- tions which would prevent the needless destruction of the small mussels. With such reasonable protective legislation, supplemented, preferably, by artificial propagation, the depleted regions generally might be aided to recuperate. Several States have in recent years enacted comprehensive mussel laws whose effective enforcement will go far to insure the perpetuation of the mussel resources and thereby the permanence of the mussel fishery and its dependent manufacturing industries. There is no question that all of the better mussel streams are capable of supporting mussel resources many times as abundant as they do now, for they did so a score or less of years ago. For each stream, therefore, it is merely a question of whether common- sense measures will be applied to restore the abundance of mussels for the benefit of all or whether they will always exist only as scattering survivals of an over zealous fishery. The conditions and the measures for protection have been fully discussed in other publications of this Bureau? and need not be enlarged upon here. It may be said, however, that there is no important stream in which the mussel resources now exist in anything like their former abundance. There have been published photographs showing fishing through the ice in the Mississippi River in the early days, where the persons are grouped closely, each one with a considerable pile of shells about the hole 2 Coker, Robert E.: The protection of fresh-water mussels, U.S, Bureau of Fisheries Document No. 793, 23 p., 2 pl. Washington, 1914. Coker, Robert E.: The utilization and preservation of fresh-water mussels. Transactions American Fisheries Society, Vol. XLVI, No. 1, New York, 1916. Smith, Hugh M.: Fresh-water mussels, Economic Circular No. 43, U. S. Bureau of Fisheries, 5 p. Washington, 1919. 110307°—21—_4. 46 BULLETIN OF THE BUREAU OF FISHERIES. through which he worked with a rake.* Such photographs could not be duplicated now, for each sheller would have to work a large area, and probably no considerable quantity of shells could be taken without more ice cutting than the value of the product would justify, even at the higher unit prices prevailing. The practice of winter shelling has, therefore, been discontinued. APPARATUS AND METHODS OF FISHERY. BAR AND CROWFOOT HOOKS. PRINCIPLE OF CAPTURE.—This method of mussel fishery is the one in most general use to-day, since it is adapted for the greatest variety of conditions, is easily operated even by the inexperienced, and the construction and maintenance involve slight expense. The method is based on the characteristic habits of fresh-water mussels, which lie habitually half embedded in the bottom, with the hinder end of the shell LL Fic. 1.—Various types of crowfoot hooks. directed against the current and slightly gaping. If a stick or hook be inserted into the opening of the shell, the mussel at once closes tightly and will hold for a long time, even while being dragged over the bottom and hauled up to the boat. Mussels are thus sometimes accidentally taken on ordinary fishhooks, while pearl fishermen working in shallow water have long employed a sharpened stick that could be inserted into the opening of the individual mussel. The more elaborate apparatus now used was first brought to the notice of the rivermen of the upper Mississippi early in the spring of 1897, and its use soon spread throughout all of the commercial shell districts. DESCRIPTION OF APPARATUS.—The crowfoot apparatus consists essentially of a bar or brail to which many short lines are attached bearing four-pronged wire hooks arranged at intervals (Pl. XXVII, fig. 1). By means of a towing line the bar is dragged above the bottom, while the hooks trail on the mussel bed in a direction parallel to the current. When a hook enters a shell opening, the mussel closes firmly upon the hook, «Smith, Hugh M.: The mussel fishery and pearl-button industry of the Mississippi River. Bulletia, U.S. Fish Commis- sion for 1898, Vol. XVIII, p. 289-314. Washington, 1899. (See plates 67 and 68.) Bu. U.S. B. F., 1917-18. PLATE XXVII. Fic. 1.—Bar and crowfoot outfit for taking mussels, consisting of john boat, two bars with crowfoot hooks, and the ‘mule’’ (lying on stern of boat). (See p. 46.) Fic. 2.—Shore outfit, consisting of cooker (at left), sorting table (center), and shells ready for sale. (See p. 60.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 47 and in consequence is dragged from the bottom. When the bar is raised after a suitable time, numerous mussels may be hanging from the hooks. The essential parts of this apparatus and the manner of its operation will be described in detail. Hooks and mode of making them.—Although the principle of the crowfoot and the general method of manufacture are the same throughout the country, there is much variety in the style and size of hooks in use on the different rivers, and even in the same camps. Some of the most popular kinds of hooks are the single-eye, the double-eye, the ring, the wrapped, the untwisted or straight-wire hook, etc. (fig. 1). Manufactured hooks are obtainable on the market, but the mussel fishermen more commonly make their own hooks, employing odd moments for this purpose with a corresponding saving in expense. The material is usually No. 11 galvanized or telephone wire. If very heavy work is to be done, a larger size, No. 9 or 10, may be taken, although hooks from the stiffer Fic. 2.—Process of making crowfoot hook. wire are more difficult to make and cause more trouble when the apparatus is fouled on the bottom; the bar may be entirely lost from a hang-up if the hooks will not straighten out before the line breaks. To make a hook one needs only a bench, an iron vise, or, preferably, an iron strap or steel plate with proper holes drilled through, a pair of pliers, and a pin or short rod for the twisting process (fig. 2). The iron strap or steel plate is usually from 6 to 8 inches long by 1.5 inches wide and 0.25 inch thick. Near one end four holes of sufficient size to admit the wire are drilled in the corners of a 0.75-inch square. ‘Two or three additional holes are drilled in the opposite end of the plate, so that it may be fastened securely to a solid block, timber, or tressel of wood, leaving the end with the four small holes free. The 48 BULLETIN OF THE BUREAU OF FISHERIES. wire is first cut into lengths of about 10.5 or 11 inches, or up to 14 inches for extra long hooks (fig. 3). The ‘‘needles” thus made are then bent into ‘‘ hairpins” or loops, with the sides parallel or nearly so. Two loops are placed diagonally into the small holes of the square and are forced down to the face of the plate, leaving just room enough for the turning pin. The twisting is done by hand and continued until there is about 1 inch of straight wire remaining in the plate. The hook is withdrawn and is complete, excepting that the ends must be cut off at even lengths and bent to the desired angle with pliers or with a piece of hollow umbrella tube. In making the single-eye hook one loop is placed half an inch in advance of the other, when they are introduced into the iron strap. In order to obtain the best results in making hooks, the holes in the plate should be kept well greased. The process is well illustrated by figure 2, while some of the various patterns of hooks made by the mussel fishermen are shown in figure 1. Fic. 3.—Stages in the process of making crowfoot hook; “‘needles,”’ “hairpins,” and nearly completed hooks. a Bar and lines.—The bar, or brail, consists of a black or galvanized iron gas pipe from 12 to 20 feet in length and with a diameter of from 0.75 to 1 inch (Pl. XXVII, fig. 1). Caps are used or wooden pegs are driven into the open ends of the pipe to keep out the water; otherwise the bar would fill with water and cause an undesirable slop when raised to the standards of the boat. The bars are occasionally supported by small wheels at the ends, to prevent the bar from disturbing the mussels before the hooks have reached them. This is generally unnecessary, because, while the boat is in motion, the bar is slightly raised from the river bottom, and only the hooks can touch the mussels. On a few rivers a wooden bar is used by some mussel fishermen, but it does not appear to be so popular or to give as satisfactory results as the iron bar. The strings or lines for carrying the hooks consist usually of soft trot-lines of No. 96 or 120 size, and are about 3 feet in length. They are attached to the bar at intervals of from 4 to 6 inches by a half-hitch knot, which is easily tied and readily loosened if a new FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 49 string is to be substituted. The soft line wears better than any other kind, especially where the bottom has much sand and gravel. Chains, or long wire links, are sometimes used in place of the cotton twine, but, although they last much longer, they are not often employed, owing to the extra weight and cost. Each line carries from two to six hooks, attached at intervals of sufficient distance to prevent successive hooks from interfering with one another. The bridle is a small-sized rope, about 0.5 inch in diameter, and is attached to the bar near the ends. It is left loose, so that it may be tied to the main rope about 2 feet above the middle of the bar. The main rope, or cable, is larger, being usually about 0.75 or 1 inch in diameter, while its length varies according to the depth of water, generally from 25 to 35 feet being required. ‘The rope is tied to the middle of the bar and to the center of the bridle. To obviate difficulty from twisting and to make the rope available when desired for other purposes, the attachment to the bridle is usually made by a swivel snap and ring. The mule.—An essential feature of the outfit with this method of fishery, under ordinary conditions, is the underwater sail or mule, as it is called. When the mussel fisherman is ready for work and the boat is over the shell bed, the bar is thrown over- board. The hooks of the bar catch in the river bottom, as well as in the shells, and thus, acting as anchors, will stop the progress of the boat downstream, unless additional power is supplied. In order to derive this power from the current, the mule is lowered into the river at the stern of the boat, to which it is connected by guide lines. Its broad surface is at right angles to the current, and the entire outfit is thus forced slowly down the river, stern first, in spite of the drag on the bottom. By means of the guide lines the mule is easily regulated to steer the boat at a desired angle over the mussel bed, or to avoid a familiar snag. There are two well-known types of mules in use on the different rivers—the common frame type and the roll mule. The frame type is rectangular in form, the outline or framework consisting of light strips or narrow boards about 3 inches wide by 0.75 of an inch thick, sometimes with a center strip extending from the middle of the bottom to a few inches above the top. In Plate X XVII, figure 1, a mule is shown lying on the stern of the boat; see also Plate XXIX, figure 1. The frame is covered with strong cloth, such as muslin, canvas, tar- paulin, gunny sacking, etc., which is left rather baggy in order to make the appliance more steady in the water; otherwise when the current strikes it at an angle it is liable to turn over, dart forward, and ‘‘kick.’’ This characteristic accounts for its technical designation as ‘‘mule.” It is connected to the boat by lines running from the four corners, excepting for those patterns having the center strip, when three connections are made with the lower corners and the top of the strip. The roll mule is not used so extensively as the frame type, but it is very popular on the Illinois and some other rivers. It consists of a piece of canvas, tarpaulin, or heavy cloth cut according to the size and shape desired. An iron rod is attached to the bottom of the mule and a wooden bar at the top. To the four corners of the cloth small lines are secured for the purpose of adjusting the mule to the boat. This form of mule has the advantage over the other in that when not needed as a sail it can be rolled up and put out of the way in the boat or used as a tent against unfavorable weather. It is 50 BULLETIN OF THE BUREAU OF FISHERIES. claimed that it is better adapted for steering the boat diagonally, and also that it does not kick. There is much variety in the form and size of the mules, which are made according to the notions of the individual fisherman or in adaptation to the condition of the river. In a shallow stream the necessary surface is obtained by making the mule long and narrow, at times about 8 feet long by 15 inches deep. For deeper waters the sizes vary according to the strength of current and the drag of the bar; 2.5 by 6 feet and 3 by 7 feet are common sizes. Boats.—The most satisfactory boat in use for the crowfoot method of fishery is the ordinary john boat, since it is inexpensive and may be made in the camps by the mussel fishermen themselves, according to their needs. Its length is from 14 to 20 feet, with a width at the center of from 3 to 4.5 feet, but it is always narrower at the ends. It has square ends, a broad, flat bottom, long rakes particularly forward, and is usually of light draft (Pl. X XVII, fig. 1). When the john boat is built particularly for this method of fishery, all unnecessary interior parts are omitted, while the needed special appliances are added. These con- sist of two perpendicular uprights or standards on each side of the boat a yard or so from the ends, a cleat at the bow, and nails or pegs at the stern. The standards are about 4 feet high and are made of light strips of wood, with notches at the top for holding the bars. When the shelling is very heavy and the bars are difficult to raise, there are added at times substandards or short strips of wood projecting outward from the stand- ards near the gunwale; in this case the bars when lifted are first placed upon the sub- standards and then transferred to the standards. At the present time nearly all of the john boats are equipped with gasoline engines of power commensurate with the size of the boats. In the Black River, Ark., some of the john boats are propelled by small stern paddle wheels operated by hand power with a vertical lever on the side of the boat. OPERATION OF THE CROWFOOT BAR.—When the john boat and all the appliances are complete for this method of fishery, the boat is either rowed or propelled by gaso- line power to the mussel bed upon which the work is to be done. After selecting the exact locality for the first haul, usually near the head of the bed, the mussel fisherman lowers a bar into the river in such a way that it will lie at right angles to the shore and drag parallel with it. The rope connecting the bar is played out until the latter is dragging freely and is then secured to the cleat. Occasionally two bars are used at the same time; the second bar being placed into the water a short distance in advance of the first and a little to one side, with a shorter rope connection. Unless the current be very strong, which is seldom the case where a good mussel bed is found, it is necessary to bring the mule into operation; and this can be so ad- justed by the guide lines as to make the outfit go very slowly or more rapidly as desired, as well as to cause the boat to sheer toward or away from the shore. After making a haul of about 100 yards the bar is ready to be raised. The method of procedure is to remove the mule from the water and then slowly draw in the connecting rope until the bar can be grasped by the hands and raised to the tops of the notched standards. ‘The other bar is put into the river, and the mule is again set. ‘The shells are then taken from the hooks and are thrown into the bottom of the boat. ‘The process is repeated until the bed has been worked over, when the boat is returned to the initial point or taken ashore, if a boatload of shells has been obtained. Bui. U.S. B. F., 1917-18. PLATE XXVIII. Fic. 1.—Barge with long crowfoot bars, employed on the Ohio River. Note rollers set on ends of bars to facilitate movement over the bottom. (See p. sr.) Fic. 2.—Lowering the crowfoot bar into the water, Ohio River. (See p. 51.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 51 Besides mussels the hooks bring up snags, small logs, and an almost unlimited variety of articles that have found their way into the river. There may be a greater or less number of larger hang-ups and other obstructions in the river, which may cause delay or theloss of a complete bar. No work is attempted in windy weather, ordinarily, on account of the trouble of handling the boat and the consequent danger of becoming entangled in the hooks or of being dragged overboard. Where no current prevails, as in Lake Pepin, the propulsion of the boat was formerly accomplished by dropping an anchor with a very long cable attached to a windlass on the boat. The boat was then propelled away to a point where the bar was dropped. Turning the windlass by hand, the boat and bar were dragged over the mussel bed. The engine power itself was regarded as too violent, as well as too expensive, for the rate of movement desired in dragging. Now, however, the shellers on Lake Pepin generally use two boats, a flat boat attached broadside against the stern of a motor boat, T fashion. In this way two or even four bars may be dragged on the bottom at the same time, employing the motive power of the engine. It is interesting to note that when this method was first brought into use in 1897, the parts of the apparatus were small and the method of employment crude. The bar was only from 4 to 8 feet long, provided with 16 or more hooks, and dragged by a rope from the stern of the boat. Two men usually operated in partnership, one man hand- ling the apparatus, while the other rowed the outfit laboriously over the mussel beds. The hand motive power was later improved by the use of a driftboard or mule. By chance it was discovered that a similar effect was had when the boat was allowed to drift broadside to the current. Although this method is still used in some places, it has not gained general favor with the mussel fishermen, probably because, when the boat is used broadside, there is more or less danger of dipping water or swamping. When the boat is used broadside, a series of cleats are placed on the gunwales of the boat in the middle portions. If a drag rope is attached to the middle cleat, the pull of the drag will be directly opposed to the current. If, however, it is desired to steer away from the shore, it is only necessary to shift the rope to another cleat, shore- ward, or channelward, as the case may be, and the resultant force of the current is in the direction desired. If there is not sufficient force in the current to move the boat fast enough, a leeboard, or mule, may be used as readily as with the ordinary fore-and-aft position of the boat. For work on a much larger scale than can be accomplished by means of the ordi- nary-sized boats there are occasionally employed heavy barges of a type illustrated in Plate XXVIII. These are used successfully on the Ohio River, near Vevay, Ind., and, though somewhat similar in construction to the usual john boat, they are much larger and more solidly built; the dimensions are, approximately, 10 by 40 feet. The barge is fitted with uprights and pulleys for handling the bars and with standards for holding them when raised. There are 4 bars 20 feet in length by 1.25 inches in diameter, to each of which are attached 76 strings, bearing 7 hooks each, thus making more than 2,000 hooks for the entire outfit. In operating this contrivance the bars at the opposite corners are lowered alternately into the river, so that as far as prac- ticable two bars are always in the water. Because of the weight and the resistance of the bars on the bottom, a very large mule is used during a good stage of water or in 52 BULLETIN OF THE BUREAU OF FISHERIES. a strong current. In the low water of summer a mule is of no avail; at this season of the year a cable 400 to 500 yards in length is used, one end of which is anchored down the river, while the other is hauled through a pulley by means of a two-horsepower gasoline engine, located near the center of the barge. The engine is also employed to assist in raising the bars from the mussel beds. The barge is towed from place to place by a small gasoline boat alongside or at the stern. By this method three men have been known to gather 3 tons of shells in a day in favorable localities in the Ohio River. ADVANTAGES AND DISADVANTAGES OF THE METHOD.—Except where snags are preva- lent, good successis had with the bar and crowfoot undera wide variety of conditions. The daily catch probably averages less than 500 pounds of marketable shells. In severely depleted regions only 100 to 200 pounds may be taken, while a half ton or more may reward the fisherman in better localities. The fishermen claim that the mussels or clams bite best in the spring of the year, on rising water, and early in the morning. A serviceable john boat could be made in 1914 at a cost of from $10 to $15; and the bars, hooks, and lines at from $5 to $6 per pair; the necessary ropes cost from $2 to $3, making a total of from $17 to $24. However, if an engine of suitable power is installed in the boat an additional amount of about $50 to $100 should be added to the above sum. On the basis of the prices of materials in r919, these costs appear approximately as follows: Boat, $23; bar, hooks, and lines, $10; ropes, $3; total, excluding boat engine, $36. ; The method has these advantages: It is inexpensive, and not necessarily laborious; it is adapted for use in deeper waters where the hand rake or the tongs can not be used successfully, and it can be employed readily by the inexperienced. The disadvantages of the crowfoot are not so obvious but are very important nevertheless: 1. The mussel beds are repeatedly dragged over by hundreds and thousands of hooks, with consequent possible injury to the mussels, especially the young. Gravid mussels, it is known, will often abort the immature spawn when disturbed. 2. Some mussels, after taking on the hooks, are pulled off while yet on the bottom with more or less injury. Experiments conducted at the Fairport station indicate that a large percentage of such mussels receive injuries from which they die. A considerable number of mussels were taken by hooks and by rakes; each set of mussels was divided into four lots, which were carefully balanced against one another in experimental ponds. After two months 38 per cent of the crowfooted mussels and only 5 per cent of the raked mussels had died. 3. The hooks take exceedingly small mussels, even down to 0.75 to 1 inch in length, which are not only useless for any economic purpose but are liable to a heightened mor- tality when thrown back into the river. The use of larger wire for the hook has been sug- gested, with a view to lessening the number of small mussels taken. There are two or three designs of patented hooks on the market, and it is claimed that they have advantages over the ordinary kinds made by the mussel fishermen. One design, invented by the late J. F. Boepple, is like the ordinary twisted-wire hook, except that the wire prongs are compressed near the tips and finally expanded to form a ball or globular tip larger than the diameter of the wire. When the ball enters the opening of the mussel, the shell closes on the compressed neck, and it is very difficult for the mussel FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 53 to fall off with the subsequent dragging on the bottom. These hooks are also slightly weighted by a wire wrapping at the lower end of the shank, and a dip in melted solder makes the entire hook less liable to untwist. Another hook, lately brought to the at- tention of the public and known as the “‘sanco-point”’ hook, has five prongs made in one piece and attached by means of a swivel to a center shaft. The tips are also glob- ular to make a so-called locking device intended to prevent the escape of the captured mussels. For both of these designs, it is claimed that the small mussels are not captured on account of the enlarged tips, and that when the ordinary-sized mussels are once caught they do not fall off the hooks, so that no injured mussels are left in the beds.* These claims remain to be effectively demonstrated; but such improvements are eminently desirable and worthy of careful test, for there is no question but that the ordinary crowfoot hook is distinctly injurious and that its use should be permitted only for a brief time, allowing opportunity for effectively improving it or displacing it altogether with other equally efficient apparatus. Meantime, mussel fishermen everywhere are urged to learn the use of other methods, for it is evident that an injurious mode of fishing will not be tolerated indefinitely. The shellers themselves will recognize the propriety of excluding from use, wherever it can be replaced, an appliance which is actually destructive of shells that are not taken or that can not be marketed when captured. Various other methods now in practical use will be described in the following pages. DIP-NET DRAG. ORIGIN OF THE METHOD.—The dip net, as used in shelling was invented and intro- duced during the spring of 1911 at Peoria, Ill. It had long been known that Peoria Lake— that part of the Illinois River which broadens into a lakelike expanse above the dam at Peoria—contained large beds of commercial mussel shells of good quality, but previous to I9II no suitable method of taking them had been devised. The various tools and ap- pliances, as the bar and crowfoot hooks, tongs, scissor forks, ete., which had been operated so successfully in other mussel rivers of the Mississippi Basin and in the major portion of this river, proved unsatisfactory in Peoria Lake. There was urgent need for some contrivance that would collect the shells in deeper water, where practically no cur- rent prevailed, and the dip net came to fill this want. It is not known who invented this appliance, but probably the idea developed by a combination of the principles of the ordinary dip net as used in fishing and the clam rake. At the present time this apparatus is used in Peoria Lake almost exclusively, none other being employed, except in places where the bottom conditions are unfavorable for the operation of the dip net. Within very recent years its use has extended to other parts of the Illinois River and to Lake Pepin. One dip net was seen on the White River of Arkansas in 1913, but it had not been put into use. The dip net is simple in construction, and in operation; it is also inexpensive and especially suited to those rivers and lakes which have soft mud bottoms free from obstructions, such as logs and hang-ups, and where there is but little or no current. DESCRIPTION OF APPARATUS.—There appears to be no definite standard or general specifications for this mechanism, and consequently there are no two alike; the black- @ Several tests made by J. B. Southall, shell expert of the Fairport station, indicate that about 30 per cent of the mussels catching on ordinary hooks are lost, while only about 15 per cent of the mussels catching on the Boepple hooks drop off. 54 BULLETIN OF THE BUREAU OF FISHERIES. smiths make them according to orders and with the material at hand. However, the various designs and patterns are very similar, the main differences being the size of the hoop and the length of the attached net. The method of operation is the same for all of them. The frame of the dip net consists of a heavy iron hoop of one piece flattened on one side. The general form, therefore, is somewhat triangular, the bottom being straight, while the two sides are curved and attached by bolts to a pole or handle 16 to 20 feet long (fig. 4 in text and Pl. XXIX, fig. 4). A large net of 2-inch mesh, made of small chain or trot-line and having a capacity of a bushel or more, is fastened to the hoop by means of chain links, and trails behind it. A short bridle attached to the two Fic. 4.—The dip net used in taking fresh-water mussels. curved sides of the hoop lead forward to a single rope secured to the bow of the boat. To withstand the strain from dragging through the water and also to support the net with a heavy load of shells, the hoop is usually made of stout wagon-tire iron, about 2 inches wide by 0.25 inch thick. The straight bottom is from 18 to 36 inches in length; the edge is bent downward and usually provided with coarse teeth 6 to 8 inches long, and at times two or three additional teeth are riveted to the curved sides, near the bottom. However,in some hoops the teeth are omitted altogether, since none are needed where the bottom of the river is composed of very soft mud. ‘The net varies in length Buu. U.S. B. F., 1917-18. PLATE XXIX. Fic. 1.—Shell tongs or scissors fork. The “mule’’ (in foreground) is not used with tongs, but in connection with crowfoot bars (in background). (See Pp. 56.) Fic. 2.—Two shell tongs (at left) and two shoulder rakes (at right). N : s . Not drift boards attached to handles of rakes. (See p. 56.) ae) a Fic. 3.—Taking mussels with the shell tongs. (Seep. 56.) Fic. 4.—Dip net and forks with shore equipment. (See P. 54.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 55 from 3 to 4.5 feet, according to the size of the hoop; when not made of chain, it is usually tarred to insure longer life. Soft twine is preferred to hard twine; the sizes used range from 36 to 96, but the latter size is the one generally employed. Power is a very important factor in the use of the dipnet. ‘The boat and engine, in fact, govern the size of the appliance, since it would be useless to work with a large dip net and very little power. Any ordinary john boat or launch, which the sheller may have, can be readily fitted up for use. Uprights or standards on the boat, as well as the mule, can be dispensed with. The engines are of the gasoline type and are from 4 to 20 horsepower. The cost of the complete dip net, including the necessary ropes, is about $5 to $7, depending upon the size. Considering the good results obtained by this method of mussel fishery, together with the durability of the apparatus, the first cost is very small, indeed. OPERATION OF Dip NET.—To operate a boat successfully, two men or a man and a boy are needed; one attends to the dip net and the steering, while the other looks after the engine and assists with the shells. When the boat is over the mussel bed, and running at full speed, the operator stands in the stern and steers with his foot or leg while manipulating the dip net with his hands. The apparatus is put into the water, usually at his right side, and when it reaches the bottom he bears down heavily on the handle. The towing line is attached, as previously mentioned, to the bridle at one end and to the bow of the boat at the other. The dip net, therefore, functions as a dredge, while the pole is handled by the fisherman in the stern in such a way as to change the angle of the net and cause it to dig into the bottom more or less deeply. Physical energy and endurance are required on the part of the operator, if the dip net is large and the power strong. If the water be rather deep or the boat very short, the angle formed by the towing line and the upright handle may be too sharp for proper manipulation. In this case a boom pole is rigged out from the bow and a longer line attached to its forward end. The teeth on the lower edge of the hoop dig up the mussels, which, due to the motion of the boat, roll into the net. Unless the net is placed at the stern in direct line of travel, the boat is retarded on one side and consequently makes a large circle over the mussel beds; this is usually desired. When the apparatus is raised after making a haul, the mud and small shells are washed out as well as can be done rapidly, and the contents are dumped into the bottom of the boat. The partner attends to the sorting out of the mussels, the trash and some of the dead shells being thrown overboard. By this method of mussel fishery two men or a man and a boy have been known to take from 1,500 to 1,800 pounds of shells in half a day in a good locality. ADVANTAGES OF THE Dip-NET METHOD.—The dip-net appliance is strongly recom- mended to the attention of mussel fishermen, as it is especially adapted for use on soft- mud bottoms and in waters which are without strong current and also where the depth is too great or the mussels too scattering for the successful operation of the rake or tongs. It may also be employed where there is a good current, providing the bottom conditions are satisfactory. The method will be at a disadvantage in very hard or gravelly bottoms or where there are numerous obstructions; in the first case the net will become overloaded with rocks, and in the second the progress will be stopped altogether. 56 BULLETIN OF THE BUREAU OF FISHERIES. It is well to point out that the meshes of the net should be of such a size as to per- mit the small shells to pass through and remain at the bottom. Some small mussels will undoubtedly be held in the net by the mud and larger shells, but these can be culled out readily and returned to the water without any material injury. It may be noted that occasionally some of the thinner-shelled mussels, such as the floater, paper-shell, etc., are pierced by the teeth of the dip net, which, of course, kills the mussel. These shells, however, are not now of any commercial value. SHOULDER RAKE. The shoulder rake can be used to advantage in comparatively swift water, espe- cially when the bottom is not too hard and is free from hang-ups such as rocks and sunken logs. The implement consists of a metal rake about 1 foot long and provided with 10 to 12 coarse teeth or curved tines, which may be about 9 inches long (Pl. XXIX, fig. 2). The rake is securely bolted to a wooden handle 15 to 20 feet long, its length being adapted to the depth of water. A basket, made of poultry-wire netting, is attached to the rake and handle in such a way as to afford a concave receptacle for the shells. A small board, about 10 inches by 2 feet, is usually fastened to the handle approximately 1 yard from the base. The method of operation is rather simple, though laborious. The boat in which the work is carried on is anchored over the mussel bed, and the rake is placed into the river at the head of the boat and slowly worked down to the stern, when it is raised to the surface. The shells are thrown into the boat. The board attached to the handle offers resistance to the current, and thus is of considerable assistance in raising the rake, as well as in driving it down- stream over the bed; it therefore has the same function of an underwater sail as the mule used in crowfooting, but the power of the current acts only upon the rake, and not upon the boat. The shoulder rake may be made from a coke fork. The tines are ctit to the desired length, heated, and bent at right angles to the handle. A Jong handle must, of course, be substituted for the short handle of the coke fork. The results of this method are generally satisfactory, if the shells are relatively abundant. Small shells inadvertently taken can be thrown back with assurance, generally, that they will live. The shoulder rake is a common implement on the Mississippi River and other streams. On the St. Francis River, Ark., it is the principal method employed in the summer and fall, while the crowfoot is chiefly used in the high water of spring. The fork, to be described later, is also used in very low water. SHELL TONGS. The shell tongs, or scissor forks, are used to some extent on the upper Mississippi, the Cumberland, the White, and some other rivers where the work can be carried on satisfactorily in rather deep water. It is possible, of course, to work between the free spaces of a series of logs or other obstructions. It is essentially a grapple, consisting of two rakes, or forks, on the ends of long handles which are pivoted together about 2 feet from the lower end, after the fashion of a pair of scissors (Pl. XXIX, figs. 1, 2, and 3). ‘The method of its operation is similar to that of the oyster tongs; the appli- ance is lowered into the water from an anchored boat, then by bearing down on the Bu. U.S. B. F., 1917-18. PLATE XOOX, Fic. 1.—The dredge ready to be lowered into the water, Black River, Ark. (See p. 57.) Fic. 2.—The dredge resting on gunwales of two boats forming the catamaran from which it is operated, (See p. 57.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 57 handles and working them together the forks are forced into the bottom and closed on the mussels. When it is closed, the handles are held together while the implement is raised hand over hand to the surface; after washing out the surplus mud and sand by a vertical motion the shells are dumped itito the boat. FORK. Occasionally a common fork is used in the smaller streams, more especially during the low water of summer. The tool generally employed for this purpose is the coal or coke fork, such as is shown in figure 5. The manner of working with the fork is similar to that of spading a garden. The operator wades into the water from the shore or from his boat and begins to fork over the mussel bed systematically. On bringing each load to the surface of the river, the sand and mud are washed from the fork by dipping it a number of times into the water; the shells are thrown into the boat, which is always kept near by. Since it is much easier to go with the current, the shellers usually work downstream, and in consequence have more or less muddy water to contend with. © The method often yields good results, but involves more or less exposure to the water. It is not particularly to be commended, since the complete digging up of the beds is detrimental to the smaller mus- sels, as well as to the bed itself; the sand and mud of the bottom, to a certain extent at least, are carried away by the current to be depos- ited lower down in the river’s course. Since the fork can be employed only in very shallow water and during warm weather, its use is consequently limited and irregular. It may be found in use during low stages on the St. Francis River, Ark., the Wabash River near Vincennes, Ind., and in various isolated localities. DREDGE. The dip net is sometimes referred to as a dredge. ‘There is also 1G. s.—The fork used a typical mussel dredge that has been in limited use in Arkansas ca an tz since 1912. While this dredge requires a greater initial outlay than — water. the simpler forms of apparatus previously described, it offers much promise as a profit- able method of taking mussels. The apparatus is well shown in the accompanying pho- tographs (Pl. XXX, figs. 1 and 2) and requires but a brief description. The dredges are of various dimensions; of the two particularly observed, one was 18 by 24 inches, the other 36 by 72 inches. The dredge may be described as composed of two heavy, long-toothed rakes with the iron handles so pivoted together scissors fashion that the two rakes when closed or brought together, form an oblong basket. Each half of the smaller dredge was 18 by 24 inches, the tines being 8 inches long and made of five-eighth inch square iron, pointed at the free ends. ‘The remainder of the basket was made of flat iron about 1 inch wide. The dredge is operated between two boats firmly attached together by cross decking at the ends, but with a suitable space left between them. In Plate XXX, the lower figure shows the larger dredge, 3 by 6 feet, spread, and held in this position by dogs on one side; it is resting across the boat. When the dredge is to be lowered, it is raised by the windlass until free of the boat, then swung around 58 BULLETIN OF THE BUREAU OF FISHERIES. by hand to a fore-and-aft position (Pl. XXX, upper figure); it is then lowered into the water by unwinding the windlass. The line from the windlass passes through a block overhead (not shown in the picture) and down to the bridle of the dredge. The two pulleys through which the bridle passes should be noted on the ends of the dredge handles. When the brake on the windlass is thrown off, the dredge falls to the bottom, and the dog releases automatically. The dredge now rests on the bottom, covering a space 3 by 6 feet, with the tines of the two ends sticking into the substratum. The first effect of turning the windlass, after taking up the slack, is to lift the ends of the handles and bring them together, thus causing the dredge to close. As the dredge closes, the tines thoroughly rake the bottom, and when completely closed every mussel and rock in the space covered, except those so small as to pass through the openings, are taken in the basket. Continued winding of the windlass brings the dredge out of the water, when it can be lowered into one of the boats and opened. All débris must be sorted out and thrown away. A small hand rake, like a flower rake, is used to clear the small stones which may have been wedged between the tines. In view of the contin- gency that the dredge may be fouled by a log or heavy stone, it is necessary to have a clearing line attached to one end of the dredge. A small windlass must be used to operate this line if the dredge is very heavy. ‘The effect of hauling on this line is to open the dredge, which may have been partly closed, and bring it up to the surface; the haul is of course lost in such a case. Heavy dredges are more effective than the light ones. The cost of the larger dredge was $65 complete, with boats and all; but there was very small expense for labor, as the work was done during the slack season and largely by the owner. The ordinary complete cost of such an outfit would be $100 or more. If the openings between the tines are wide enough, the small shells will not be re- moved from the bottom. Comparing the dredge with the crowfoot drag, it may be noted that the latter takes mussels by chance and that repeated dragging over the same bottom is necessary to make an approximately clean catch, while the former makes a clean haul of only the mussels large enough to be taken. It will be seen, therefore, that the crowfoot apparatus, although less effective over a given small portion of bottom, is actually more destructive to the young mussels. An entirely new form of dredge has recently been invented, which is operated by power and brings the mussels continuously from the bottom by means of an endless chain and buckets. No detailed description can be given at this time. LOCAL MODIFICATIONS OF METHODS. Various other forms of apparatus have been devised at different times and put into temporary use, but none of them seems to have won a place in the established methods of fishery. There are many local variations of the typical methods described, but it is not practicable to describe them all. Two special modifications of the use of the coke fork and of the basket rake were thus described as used in the James River in 191374 The mussels [at Riverside, S. Dak.] were gathered with a coke or coal fork, having a piece of 2 by 4 lumber fastened to the handle, the length of this piece being according to the depth of the river. This fisherman had a novel way of anchoring his boat. At each end of the boat a hole was bored through the bottom large enough to insert a piece of 1.5-inch pipe, makinga water-tight joint. These perpendicular a Coker, R. E., and Southall, J. B.: Mussel resources in tributaries of the upper Missouri River. Appendix IV, Report, U.S. Commissioner of Fisheries for 1914, 17 p., t pl., map. Washington, rors. But. U.S. B. F., 1917-18. Prate XXXII. Fic. 1.—Mussel boats on Rock River Pool at Government locks, a few miles above the mouth of the river. The mussels are taken by hand while wading in the shoal water below the dam. (See p. 59.) Fic. 2.—The mussels when taken are put into a small flatboat and conveyed to the dam, where they are transferred to another boat above the dam, as illustrated. In the second boat the mussels are taken to camp for cooking out. This is the scene of one ot the most extensive mussel fisheries where shells are taken by hand, (See p. so.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 59 pipes, rising to the level of the gunwales of the boat, served as sockets or sleeves, through which a long iron rod could be shoved into the soft mud bottom of the river. By anchoring in this way the boat was kept abreast of the current, while the fisherman used the sides as a fulcrum for the handle of the fork. After gathering all the mussels possible within reach he would pull up the rods, let the boat drift down- stream a suitable distance, or beyond the portion of river just worked, and then anchor and resume operations as before. At Milltown, mussels were gathered by means of the basket-rake dragged by a power-boat. The rake was peculiar in being without teeth but having a square brail made of 0.25 by 1.5 inch flat iron, to which was fastened a wire basket of 1-inch mesh. With each boat was a crew of four men, three to work with the rakes and one to operate the engine. One dragged the rake at the rear of the boat, while the other two worked at the sides. In this manner a strip of the river bottom 6 feet wide was thoroughly scraped. SUMMARY OF METHODS OF FISHERY. The principal forms of apparatus are the crowfoot bar, the dip net, the shoulder rake (or basket rake), the forks, and the dredge. A considerable quantity of shells, about 500 tons each year, are taken by hand. (Cf. Pl. XXXI.) The statistical reports previ- ously cited (p. 39 above) show in detail the quantities of mussels taken by theseveral forms of apparatus. From these reports the percentages of the total of 51,571 tons of shells taken in the territory covered, as credited to the several forms of apparatus, may be computed and stated as follows: _ Percent. Per cent. Growioot. peapienctar cn icetrco Stir ais sols afer wlaleie PONOMIM DV Mets aaje crass cartier aeree greteie eee een 3.3 Forks. .... 00.500. .e eee eee c eee eee EO. FWEDrea pers Lice inane vein ence database the ein att 1.2 Mongatwe: evens. S62 BUSA, 1, MAREE 7 SP Rakesx).. deus. wloageniss skh’: « .saskon- < I.2 Mandsiier..cenie brig whtosm ask cre@tgas 503 = 99-3 . SHORE EQUIPMENT AND PROCESSES. It is customary for the shellers to establish camps alongshore. Sometimes the camps are individual and occupied by one sheller with his family; in other cases a sort of village camp is found where a dozen or more families of shellers are grouped. The selection of a site is governed, first, by the proximity of a good shell bed; next by the convenience to wood and shade. Rude frame buildings may be constructed, or tents may be used. A very common form of dwelling is the house-boat, or “shanty-boat,”’ as it is generally termed (Pl. XXXII, fig. 2). There are many different forms and sizes of shanty-boats, to suit the needs and ideas of the fishermen; popular sizes are about ro by 35 feet and 12 by 40 feet. With such boats it is a simple matter for the sheller to move from place to place, according to the requirements of the fishery. One or two small flatboats and usually a larger boat with a small gasoline engine are almost always employed, whether or not the house-boat is used. After bringing the mussels ashore the soft parts must be removed. Where pearling is the exclusive object, each mussel may be opened with a knife inserted between the valves of the shell, so as to sever the adductor muscles; the meat is then cut out and examined for pearls. This may be done while wading in the river and the meat and shell thrown away at once. Such a process is entirely too slow and tedious for preparing shells for market; hence the cooking out process is exclusively employed in the shell 60 BULLETIN OF THE BUREAU OF FISHERIES. fishery. The man may fish during the forenoon and cook out in the afternoon; in some cases the wife or children of the sheller attend to the cooking out, while the sheller continues the fishing operations. The cooker consists of a vat about 5 feet long by 2 feet wide and from 12 to 18 inches deep (Pl. XXVII, fig. 2 extreme left, and Pl. XXIX, fig. 4). The frame may be of wood and the bottom of sheet iron or stovepipe iron, brought up a few inches over the lower edges of the wood to protect it from the fire. The bottom of the cooker is usually made to slope upward at one end in order to facilitate the forking out of the shells. The vat is set over a trench or ditched-out furnace, the back part of which is fitted with a couple of joints of stovepipe or smokestack of some kind to furnish the necessary draft; driftwood may serve as fuel. When the cooker is filled with mussels, a small amount of water is added, and the whole is covered with burlap or gunny sacks. ‘The fire is started in the furnace and continued until steam is being given off in quantities sufficient to kill the mussels, so that they will open readily. The process may take about 20 minutes or longer. If the mussel camp is situated near a factory or some establishment from which steam can be obtained at a reasonable price, there is a great saving in time and trouble by making a direct steam-pipe connection between the boiler and the cooker. The shells are prepared in the same way, but instead of applying heat beneath the cooker the steam is admitted directly into the container. The shells are removed with a fork and thrown on the sorting table which is about 3 feet high and of sufficient width and length to hold at least one-half of the contents of the cooker. The mussels must be handled separately, picking or shaking out the meats, which are put to-one side for later examination for pearls and slugs, while the shells are thrown into heaps on the ground or into small bins (Pls. XXVII, fig. 2, and XXXII, fig. 1). When all the shells have been cleaned, the water, or soup, in the cooker is carefully strained through a small mesh screen of wire netting in order to recover any pearls or slugs which may have become disengaged from the meats during the cooking-out process. It is said that pearls which have lain on the hot metal bottom for any length of time are permanently injured. The size of the screen is usually about 1 foot square. Most of the pearls are found in the meats, which must be examined one by one. The pearls are not always visible, but are found by slipping the meats through the fingers. Small pearls are sometimes recovered by allowing the meats to rot in kegs or half-barrels. When reduced to a pulp, the mass is rubbed through a fine-mesh sieve, the pearls and slugs being retained on the sieve. Many mussels are cooked out merely with the hope of finding pearls. The non- commercial shells must be thrown aside, but there is no general practice of classification of the salable shells. Often this is done by throwing shells of a certain quality, such as niggerheads, pimple-backs, etc., into one pile and blue-points, washboards, and miscel- laneous shells into another. This is usually an advantage to the sheller, since he may obtain an advanced price on the best grade shells; yet the practice of buying the river run at one price is still very common. Most of the shellers do sort out the yellow sand- shells, since these command a price several times higher than the others; but even this is not always done, and thus a good many yellow sand-shells are received at the factories along with other shells. These, of course, are sorted out at the factory, and resold to Buu. U.S. B. F., 1917-18. _ PLate XXXII. Fic. 1.—Sorting table and heap of shells on river bank. (See p. 60.) Fic. 2.—Barges loaded with shells and two shellers’ house-boats, in Arkansas. (See p. 59.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 61 the export trade. It seems to be mere shortsightedness on the part of many shellers that prevents the effective classifying of materials before sale. The shells are sometimes sold to local buyers or local factories; more often they are bought by traveling buyers who may be agents of the larger factories or professional shell merchants who buy in large quantities and sell as they find the most favorable market. Some buyers maintain large power boats and barges (Pl. XXXIII), which may travel up and down the river, prepared to load the shells and convey them to a convenient manufacturing or shipping point. The greater part of the shells are shipped by rail in car shipments from the nearest freight station. At times and in some places it has been the practice for shellers to operate small cutting plants, but the scheme has not always worked well in practice. Each is a profession in itself, and the best cutting is done by those who are practiced in cutting and shop management and who can keep well advised as to the market demand for the various sizes and qualities of blanks. ELEMENTS OF WASTE. CULLS. The piles of culls are usually not large in proportion to the heaps of economic shells, but they include a good many kinds of shells, worthless because of excessive thinness, undesirable color, spotting, or other evident defect. Among such culls are the paper- shells, the pink heel-splitter, black sand-shells of pink nacre, purple warty-backs, and often the purple or salmon-colored elephant’s ears and spikes. The last-mentioned can be used for making smoked-pearl buttons, although they are not usually in demand. It sometimes happens, therefore, that at the close of the season a buyer will take the usable colored shells at a reduced price. The tendency of shellers to throw in all off- colored shells has given unfortunate discouragement to this practice. MEATS. In connection with the shells collected each year, there are taken some 10,000 tons of wet meats for which there is no appropriateuse. Small quantities are sold locally for use as fish bait with trot-lines, or hoop nets, or as food for poultry or pigs. The fresh meats, after being allowed to sour in the sun, are considered particularly good for these purposes; but generally only a small proportion of the meats has been so used. It is often a serious question in the mussel camp to make proper disposal of this material. The meats that can not be sold locally are often dumped into the river, buried in the ground, or put into a “‘rot box.” The throwing of meats into the river in large quantities becomes objectionable when those that are not eaten by the fish and turtles rise to the surface in a state of decomposition and are washed ashore to cause an offensive stench in the neighborhood of the camps. The meats, when dried in the sun or by the use of artificial heat, can be ground to make a fine meal, in which condition they appear to keep indefinitely. For the purpose of sun-drying they are spread on frames made of coarse-mesh wire screen so arranged that the air can circulate freely between them. In dry, sunny weather the meats can be dried in from 30 to 72 hours to about five-eighths of their wet weight. When so. dried, they can be ground in a coffee mill or similar machine; but the foot part becomes exceedingly hard and tough when partially dried and rapidly wears out the mill. Since the meats are usually too large to feed into the coffee mill whole, they should be reduced 110307°—21——_5 62 BULLETIN OF THE BUREAU OF FISHERIES. by pounding or, better, by chopping in a meat chopper before drying. Sand should be avoided in the whole process, as it will damage the grinder. If the foot is to be ground, the whole meats must be dried by artificial heat to about one-seventh of the wet weight. This entails some additional expense for fuel and ovens. The loss of nutritive substance in discarding the foot is not great; but it is obvious that practical difficulties are encoun- tered in separating the tough from the soft parts before grinding. The whole trouble arising from the toughness of the foot is obviated by putting the wet meats through a sausage grinder before drying and then regrinding the dried masses of meat in any suitable mill; the product thus obtained is not a fine meal, but a coarsely granular material, practically dust free and very suitable in form for use as food for poultry or fish. At the Fairport station the ground meats have been found to be very acceptable to chickens when moistened to make a mash and mixed with grain. Experiments made by the Bureau of Animal Industry show that the dried mussel meats are a suitable food for chickens, having about the same value as fish meal. To obtain like results, a slightly larger quantity of these substances than of meals made from the red meats must be used. For some years fresh mussel meats have been used as a food for fish at the station, and it has been found best to allow them to sour a little before feeding them. The ground dry meats have also been used in feeding small fish in aquaria and in ponds, and they have proven a very satisfactory food material. The ground mussel meats have recently appeared upon the market in the form of feed for poultry and fish. The meats of the mussels could, perhaps, be used for human food if they were collected under sanitary conditions and properly prepared. This question should be subjected to experiment; but it would be obviously impossible to consider with reference to human food the use of meats saved as a by-product of the shell fishery under the conditions now prevailing. Analysis of the mussel meats made for this Bureau by the Bureau of Chemistry indi- cates a very desirable content of protein, glycogen, phosphoric acid, and lime, if the meats are considered with reference to their use as a food for poultry or fish. Approximately the percentages are: Protein, 44 per cent; glycogen, 9 per cent; phosphoric acid, 9 per cent; and lime, 8 per cent. An analysis in detail of meats of fresh-water mussels from the Mississippi River is stated in the following table: ANALYSIS OF Dry MussEeL MEaTS.2 Per cent. Water, at 65° in vacuole). ee DD ee a ee eee ee 7-59 Btheriextracted!. .:ostz. aes aie. heismstls -modie. sie yeast -blee-ge- dete tees -faelh aise 2. 84 Total, nitrogen... |. p05 memeeiege gets melee nd peeysee hes mepserer sty: geet of! serene’ yo vetine ete ges Hh yaa Protein (NU O:25) oo gece slept ogo sep nin esol ype oan oie ictal alee alo coe aka ceria else geod cae oie cio on 44.44 Gly CORE swe nrsi aaioe siteto sie ete mci mtn rps sols o'r watetcta aualale oe ete) e totaal ieials tel scin maa Lae easton iole oie 9. 35 Undetermined (nonnitrogenous organic material). ......... 02.02 e eee eee eee teen eee eee ees 13. 02 DSW eos w coa cog: aseiend aia bis oie Sula ayeimioleveeleis!s:svayes ate sub e shu latedatalelstsiatelallele is/iens jetehegetaeteds/aRaite ede AdaaNele RMS lates 22. 76 Ash content: Phosphoric acid, PlOg. 0.0.2.2. c cece ee eee eee cece cence eens t ete e eee de renee seca 39. 31 Lie). CaO. 02 TIMI MAIN GES ARLROO Me REE. SALE INS SSIS SIR YS Ne aes 34-71 Silica} SiQs. 914 -saddasvs -varciies seth ik. apeelt feoge dost -yieer? -sigisois, $9,040,029 c) «+.-| $16, 233, 198 Fresh-water pearl. . . rt 1,176, 285 37359, 167 € 6} 4,879,844 Ocean pearl......... T, 951, 558 Z) 511, 107 E 2, 489, 364 Metal. .. 0.2.50... 887, 521 1,312, 741 c 763, 287 Vegetable ivory. .... 1,144,677 1,305, 766 ce 2, 885, 503 — igadadochehohs 468, 121 766, oor by sj been tS asec pocasonoseee 0 a ae See 137) 401 124) 454 c ‘ aero 2 MWothers ©. 29.0 i2ec. cade 701, 810 660, 703 ) alae |e aoe aes, 266 Button blanks made for sale 656,036 I 916, 003 c By SEXP ANTS. abeh eek - All other products. Hoa ocee ceale as sees sae 9 1,177,737 c 1077 OOF {ch cacivtee a edie DO OR oe ROC) J Ce aE ee ore 7,695,910 | © x1, 133,769 | $22, 708, 065 |............ 420, 791, 985 @ Fresh water only. > Exclusive of buttons to the value of more than $1,000,000, made in each year 1904 and 1909 by establishments engaged pri- marily in the manufacture of other products. ¢ Not classified, 4 The product of Iowa and Illinois in 1897 was $243,655 and in 1898 $252,570 (Smith). ¢Some of the materials from which buttons are made, in addition to those indicated in the table, are brass, composition (clay, etc.), wood, glass, gold, hoof, iron, ivory, leather, paperboard, porcelain, silver, steel, and also, in some cases, skim milk (casein), animal] blood, and probably bakelite. f Probably fresh-water pearl chiefly. 9 Partly fresh-water pearl products. h Includes blanks, or molds, snap fasteners, and all other products in amount, $4,558,787. The census report of rgoo states: ‘“‘In 1890 there was not a single fresh-water pearl button made in the United States. In 1900 the making of these buttons constituted the second most important branch of the button industry.”” Yet, at the next census, only five years later, the fresh-water pearl buttons are found not only in the first rank, but actually exceeding in value the combined product of the two next highest—ocean pearl and vegetable ivory.” It would be of value to compare the production in gross of buttons and the price per gross during the years from 1899 to 1909. Unfortunately, the census report gives no classification except for the years 1899, 1904, and 1914; but the table following sup- plies the blanks by computation, the basis for each computation being explained in footnotes . @ It is well known that for several reasons there wasa temporary decline of button manufacture between rg09 and r9r2. > The figures for 1914 are not quite representative for the fresh-water industry, since the blanks and by-products aggregated at the bottom of the table are probably principally fresh-water products, as may be inferred from the total for that industry given by the census report in another place as $4,370,000. 68 BULLETIN OF THE BUREAU OF FISHERIES. All kinds. Fresh-water pearl. Ocean pearl. Year. Source. Million | Price per| Million | Price per} Million | Price per gross. gross. gross. gross. gross. gross, (Oye e ice ceccrcns WD rele oshnichdre 4-3 $o. 27 4.0 $o. 48 II. 4 30 1-7 87 €30.0 SbagRalkaart > sar dkcph tsps © 126.0 POZsON ae roe cose] as cele serete 21-7 225 45 5st @ Obtained by dividing census value of $4,217,000 by assumed average price of 30 cents per gross. > None, ¢ Unknown. 4 Obtained by dividing census value of $22,708,000 (aggregate), reduced to $20,000,000 to allow for value of waste products, by assumed average price per gross of 30 cents. This average price taken from censuses of 1900 and 190s. € Unofficial estimate. / From statistical survey of the fresh-water mussel industry conducted in 1913 by the Bureau of Fisheries. IMPORTS AND EXPORTS OF BUTTONS. The history of imports of buttons of all kinds since 1891 has an interest in connec- tion with the domestic industries and is shown in an accompanying table. The exports of domestic buttons are also shown for the few years for which they have been sepa- rately shown in the schedule. (See table below.) We find, first, a substantial decline in imports approximately coincident with the inauguration of the fresh-water pearl industry, but evidently not related thereto, this decline being attributable to the financial strin- gency of 1892-1894; second, a substantial recovery of import trade in 1895 and 1896; third, a marked decline in imports coincident with the rapid expansion of the fresh- water pearl-button industry in 1897 and the following years; fourth, a general slow rise in the amount of importations, beginning about 1900, although never, until 1913, rising to more than about one-half of the importations of 1891. Nevertheless, the difference in value between the imports of 1891 and 1910 is not at all commensurate with the output of the fresh-water pearl industry. In the later years there is not a wide difference between the value of imports and exports. Imports of pearl buttons have never been of considerable value, except about 1896, 1903, and 1917. Imports AND Exports OF BuTTONS.@ Imports (for Imports (for oa ii gensuriptics) 4 4 coneimption) ports, a! of pearl an: . ‘mports, a of pearl an: ie coca kinds Exports, all | shell buttons, Spay coding kinds pee all | : including imports from imports from Philippines. Philippines. $2,096, 000 $100, 000 || 1905....-..eeee eens $866, c00 $172, 000 I, 317,000 292,000 || 1906....-...2eeeee 873,000 134, 000 I, 410, 000 4757000! |||) 'ZO07. ass we eciecicins cle 936,000 164, 000 465, 000 38,000 || 1908.....ceseeewe 653,000 93, 000 I,07I, 000 376,000 || 1909..... oe waite ae 767, 000 87, 000 I, 424, 000 B57, OOO |[ATHO ccc ccs cucases 1,056, 000 107,000 950, 000 259) 000,|). TOLL. b+. veweeswens 762, 000 100, 000 436, 000 57 OOO PLOL Sa lalaistel ste aicie'sicisiei I, 130,000 71,000 451, 000 24,000 |} TOL eco sv cvecaee es 1,856, 000 137,000 593) 000 BGS OOS EN XGLA wise « wivie dim cre ic'vies 2,082, 000 253,000 551,000 |. 36.000: I) LOLS. < aira ole - waa pee 1,005,000 280, 000 954) 000 424,000 || 1916.......0cc ence 789, 000 546, 000 I, 190,000 W4GO SOOO! ||" TOL ere cia cisia wclnivnicls I, 207, 000 1,058, coo 893, 000 SEPLOOG Hl LOLS sre aie rateinisldiemaniele 1,276,000 914,000 2 From Government reports and information furnished by the U. S. Bureau of Foreign and Domestic Commerce, but in each case reduced to the nearest even thousand. b Imported buttons of high price are principally glass, pearl, and metal; of medium price, Philippine pear] (small quality); of low price, agate, bone, and nickel bar and recently Japanese pearl. ¢ Including pearl buttons to the value of $600,666. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 69 When the imports are added to the domestic production we observe the enormous increase (in consumption) from about $9,000,000 to about $23,000,000, during the course of 20 years, an increase entirely disproportionate to the growth in population during the same period. The per capita consumption of buttons grew from about 37 in 1891 to about 106 in 1910.% This magnified consumption can not be attributed simply to increased prosperity or to growing extravagance. It is a matter of common experience that where relatively expensive buttons were formerly hoarded and used again and again, the modern cheap- ness of good material has lead to the general abandonment of this laborious practice. The real significance of the fresh-water pearl industry is that it has, by its direct and indirect effects, made good buttons low in price and more generally used; as, indeed, would be the result in any industry that found a new and abundant resource to yield a quality of product formerly obtainable only from less available and more expensive materials. During the last few years there has been increasing activity in various branches of the button industry, notably in vegetable ivory and celluloid, but fresh-water pearl still ranks as the paramount material used by button manufacturers. SUMMARY OF ECONOMIC. EFFECTS. We may thus summarize the broader effects attributable in large part, though we would not say exclusively, to the development of the fresh-water mussel industry. Although affording employment to many wage earners and giving occasion for an im- portant fishery, it has not caused a material diminution, if any, in the output of any other branch of the button industry. The fresh-water pearl product alone is now greater than the entire output of the button industry in 1890; but at the same time the product of other branches of the industry is greater than in 1890. The fresh-water product is simply an addition by so much to the available wealth of the country. It has made a good button almost universally available, so that the total consumption has been greatly augmented. The economic rating of this industry will always depend in considerable measure upon its supplying a good product at a relatively low price; and this con- dition will be maintained only by preventing the depletion of the native resources and by promoting economy in manufacture. DEVELOPMENT OF MODERN METHODS. In the early stages of the industry the making of buttons was accomplished largely by hand machinery, so-called. The shells were held against the revolving saws by hand while the blanks were being cut out. Each blank was held with the finger against a revolving emery wheel, first to be backed, or ground to a smooth surface, and next to be turned or faced to a proper form with the central depression worked out. Then the blanks individually were placed in chucks for the drilling of two orfour holes. Only the final polish was administered to the buttons in bulk. Sorting and carding was, of course, done by hand. @ These figures are based on the computation that the $4,216,000 worth of buttons of 1890 represented, at 30 cents per gross, 14,000,000 gTOSS, OF 2,000,000,000 buttons, while the output in 1910 is computed in the same way as 9,500,000,000. _Importations are added and exportations deducted. 7O BULLETIN OF THE BUREAU OF FISHERIES. Most of these processes have become obsolete in the United States with the devel- opment of modern machinery. It is, indeed, to the automatic machinery that the industry owes its present relative importance. The old process of sawing remains prac- tically unchanged, but the grinding, facing, and drilling, the principal features of button making, are accomplished by automatic machinery. For the three processes either two or three machines are used; in some cases a grinder attachment to the double auto- matic makes it possible to combine the entire process in one machine. The sorting, or grading, of the buttons requires nice judgment and must still re- main a hand process; but recently a machine has been introduced for the attachment of buttons to the cards. Not only the departures in the mechanical equipment but the improved efficiency of labor and better shop management are combining to increase the output and to promote economy of production with better quality and uniformity of product. These, and such other present-day features as the utilization of waste materials and the intro- duction of sanitary devices, will undoubtedly be more generally and effectively applied in the future. PROCESSES OF MANUFACTURE. The description of the general process of button making as given below is essen- tially that of the average modern plant, although in each factory characteristic modifi- cations of method are encountered. PREPARATION OF SHELLS. STORAGE.—When the mussel shells are received by barge or freight car, they are hauled to large covered or exposed storage bins at the factories, to be kept until ready for use. A rough sorting is often done at this stage, so that each bin will contain shells of a relatively uniform size and quality. There is no apparent deterioration of the quality of shells if protected from the weather. If not so protected, they are liable in time to lose the luster and become lifeless or chalky. The exterior of the shell is most readily affected once the horny covering is worn or scaled off. For this reason shells which have been long exposed on the banks or “‘dead” shells from the rivers are consid- ered undesirable. CLASSIFYING—When the shells are taken from the bins for use, they are first sorted by hand according to species or quality of material, if this has not previously been done, and are then classified as to size. The latter process is accomplished by a machine called a classifier, which, though larger, is similar in principle to the classi- fier used for blanks (Pl. XX XVIII, fig. 1). The shells are put into a large hopper, from which automatically they are fed slowly onto an endless belt leading to the classifier, which consists, primarily, of two hollow metal rollers about 6 inches in diameter and 8 feet in length. By falling between the two rollers the shells are to be separated, roughly, into four or five grades, according to size and thickness. To this end the rollers are set with an incline and are not quite parallel with each other, being more widely sepa- rated at the lower ends. As they revolve outwardly the shells slip, or roll down the incline to a point where the opening between the rollers permits them to fall into one of a series of buckets placed below. The smaller and thinner shells are found in the buckets nearest the head. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 71 The classification of shells by size and character has obvious advantages for adap- tation of material to particular grades of buttons and to uniform speeds of machinery. Soaxinc.—After classifying, the shells are placed into large metal tanks or vats, each holding about a ton, or sometimes into barrels. The containers are then filled with water, in which the shells are allowed to soak for a week or more. ‘The process is intended to soften the material, which would otherwise be too hard on the saws, as well as so brittle as to chip and yield blanks with rough edges. The smaller sizes of shell may simply be put into damp cellars, sprinkled, and covered with wet cloths. Curiously enough, there is a difference of opinion among manufacturers as to the merits of this treatment, the contrary view being held by some that the material works better if not soaked, but simply sprinkled, before cutting. The custom of soaking is one that is generally desired by the cutters, however, since it is held to lessen the labor of resharpening the saws. An obvious but probably unnecessary disadvantage of the soaking process consists in the fact that the shells as marketed have small bits of meat attached, so that after soaking the water may become more or less foul. In handling the shells scooped from the vats there is the possibility of contamination in the event of abrasions upon the hand. ‘This is undoubtedly the cause of an infection to which cutters are more or less liable. The addition of an antiseptic compound to the water of the vats should be a universal feature of this process. SAWING OR BLANK CUTTING. THE MAcCHINE.—The cutting machine is essentially a lathe fitted, on the one hand, with a tubular saw of the necessary diameter to obtain the required size of button, and, on the other hand, with a wooden plug and a ratchet handle or lever for gradually forcing the rough shell against the rapidly rotating saw (Pl. XXXV, fig. 3). The shell is held in position either by tongs or by the hand protected with a mitten. During the cutting process small jets of water are directed against the shell to keep it, as well as the saw, cool and also to prevent the production of troublesome and injurious dust (P]. XXXVIII, fig. 2). Successive blanks as they are cut are crowded through the tubular saw to fall into a receptacle below. When the desired blanks are removed, the shell is thrown into a bucket or box to be subsequently dumped upon the shell heap, unless the shell is to be passed to another machine for a second cutting of smaller and thinner blanks. The cutting machine, including the saw, is about the only one of the many used in button factories on which no radical improvements have been made since it was first introduced and adapted to the fresh-water shells. The original invention may or may not be the best solution of the problem, but as yet none of the machines put forth as improvements has earned an established place in manufacture. The saws must be made of specially hardened steel, and are obtainable from only a few shops. When received from the factory, each saw is simply a rolled cylinder tapering at one end, being without teeth at this stage (PI. XX XVII, fig. 1). They are made of different tempers as extra-hard, hard, regular, soft, and very soft, for adapta- tion to the varied texture of the material to be cut. The sizes of the buttons are determined by the inner diameter of the cutting end of the saw, and the unit of button measure is one-fortieth of an inch, called a line. But- tons from fresh-water shells vary in size from 14 to 4o lines (from about one-third of an 72 BULLETIN OF THE BUREAU OF FISHERIES. inch to 1 inch). Buttons of ocean pearl are sometimes made as small as one-fourth of an inch, and the same size of fresh-water buttons is made in novelty works as well as the larger sizes up to 60 lines (1.5 inches) or larger. The button-cutting machine is equipped with a three-step cone pulley so that the speed may be adapted to the shell and line to be cut. For blanks of 14 to 20 lines a speed of over 400 revolutions per minute may be used, while about 275 revolutions would be used for 22 to 36 lines. The largest sizes, 36 to 60 lines, would be cut with a speed of only about 180 revolutions per minute. The speed will, of course, be adapted some- what to the shell and to the whim of the individual operator, The cost of a cutting machine installed was estimated four years ago at about $24. (Seepage 44.) The number of machines operated in one plant varies from three or four in small blank-cutting shops to one hundred or more in larger factories. DETACHED CuTTING PLANTs.—While button factories commonly include cutting rooms (Pl. XXXVIII, fig. 2), where the blanks, or buttons in the rough, are cut from the shells, there are yet a good many establishments devoted exclusively to the finishing and grading. Insuch cases the cutting is done in detached cutting plants (Pl. XXXIV, fig. 2), which may be located at convenient points in different States and from which the blanks may be shipped to the factory. There are also numbers of independent cutting plants, or button shops, which may be more or less portable. The owners of such plants take the shells from the river or buy them and cut out the blanks, which are then sold to the manufacturers of buttons. It has frequently occurred that when a new region of abundance of commercial shells has been discovered numerous small cutting plants have sprung up along the banks or on house-boats. A single fisher- man may purchase and install a single machine and small gasoline engine to cut the shells that he and his family take, or a number of machines may. be installed, labor employed, and the product of other fishermen purchased. In a few cases the cutting plants are cooperative, a number of shellers operating each a particular machine and cutting and marketing his own blanks. The blanks may be sold to the owner of the machines or in the open market. Most manufacturers purchase blanks when it is more profitable to do so than to produce them in the factory; but generally a manufacturer prefers to produce his own blanks, since greater care can be exercised in proper cutting. There are cases where the independent cutting plants are particularly to be recommended, as where the shells are too scattering for convenient shipment in carload lots, or, in remote localities, whence the freight charges on the bulky shells are practically prohibitive. In some streams it appears that the best solution of the marketing problem would be had by the operation of small cutting plants on house-boats, which can be floated down the river, cutting the shells as they are found and throwing the waste shell back into the river. The blanks can be shipped from time to time from convenient points. This plan has also its advantages where the shells are abundant but so spotted or stained that the proportion of good shell to waste is relatively low. It is of interest to note that the freight charges paid for transportation of shells and blanks in 1912 were reported at $131,000. WorK AND WAGE OF CUTTER.—The cutters are men. Each cutter is ordinarily expected to provide himself with a few tools, such as three to five saw spuns, a button- cutter’s hammer, shell tongs, saw, files, and hose and bibb. This equipment may be FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 73 purchased either from the factory or elsewhere; its cost in prewar times was about $2 and approximately $4 in 1919. The button cutter, having had his shells weighed out to him and having received his saw, proceeds to fit the saw into the spun, or saw holder, and then sets the spun in the chuck fixed in the machine, which is thrown into operation to test if the saw is set perfectly true. If not found to be running true, adjustment is effected by tapping, or by refitting. It may require several minutes, at least, to obtain a correct adjustment. With a slender three-cornered file the teeth (11 to 20 or more in number, according to the size of blank to be cut) are then cut into the saw, after which the teeth are set, and the machine is ready for practical work. (See Pl. XXXVII, fig. 1.) In quantity of pro- duction, quality, and economy, as will be more fully shown later, much depends upon the skill and interest of the cutter, as well as upon the good judgment of the manu- facturer in the purchase and assignment of material. Roughly speaking, a cutter may use about roo pounds of shell per day, cutting 25 gross or more of blanks. The number of pounds of shell required to produce a gross of buttons varies with the line, the character of the shell, the skill and interest of the cutter, and with the care of the management. Once a week the cutters take their blanks to the foreman and are paid at a given price per gross. There is usually some system by which the cutter is held responsible for excessive waste of shell. The wages of cutters vary widely, according to the skill and regularity of the individual. Since the number of gross is computed from the weight of the blanks in bulk, it is customary to give the blanks a preliminary shaking over sieves adapted to the size to be used. The openings in the sieve are just a little smaller than the blanks, so that, not only the chips and dust are removed, but also such imperfect blanks as have one diameter less than that of the opening. Representative prices in 1914 and in 1919 (figures in parentheses) were as follows: 6 ( 734) cents per gross of 14 lines. ro (11%) cents per gross of 24 lines. 6( 8 ) cents per gross of 16 lines. 13 (15) cents per gross of 30 lines. 7 ( 8%) cents per gross of 18 lines. 17 (18 ) cents per gross of 36 lines. 8 ( 9!) cents per gross of 20 lines. 21 (26 ) cents per gross of 4o lines. 9 (1034) cents per gross of 22 lines. At these prices a cutter could earn from $10 to $20 per week in 1914, or from $12 to $35 in 1919. A small bonus may be paid for full-time work. PRODUCTION OF BLANKS.—With good cutting of niggerhead shells 100 pounds of shell will yield 12 to 14 pounds of blanks, but the production is usually much lower, often only about 9 pounds. A fine grade of muckets from Lake Pepin, being light and of comparatively uniform thickness, will yield 20 pounds of blanks per roo pounds of shell. The table following prepared by the shell expert of the Fairport station is of interest as illustrating how the different species and sizes of shells may be adapted for different lines of buttons. The cutting practice in any plant will, however, be adapted to varying market demands, rather than to any theoretically ideal scheme for the most effective use of the shell. 74 BULLETIN OF THE BUREAU OF FISHERIES. SPECIES AND Sizes oF SHELL THat May BE ADAPTED FOR DIFFERENT LINES OF BUTTONS. Lines of buttons. Common name. Species. Remarks. Small | Medium | Large shells, shells. shells. Lines. Lines, Lines. Murclketinhs ah Fie gye fore siste vg aie ch oda Lampsilis ligamentina............. 14-18 14-22 14-30 Yellow sand-shell..............0005 Lampsilis anodontoides........... 14-16 14-20 14-24 | This species usually ex- ported for novelty work, Slough sand-shell..........-6..0005 Lampsilis fallaciosa., ..........60- 14-20 14-20 14-20 SEC CU Natron kteciejatseisatinietehi ein Lampsilis luteola............+.000- 14-20 14-20 14-20 | Good shell only in cer- tain regions, WASH DOGNGS Vis ioivin ciciste'cthioniale steldimtastels Quadrula heros:......5.....0.ece0s 14-20 14724 14-40 | Often much waste on account of spotting. PLELEE-TIA LE. sis Mieieisacs ceielciesie ce ces ola Quadrula undulata 14-20 14-24 14-30 Niggerhead ..| Quadrula ebenus..... 14-16 14-20 14-24 |) These species yield a Maple-leaf....... ..| Quadrula lachrymosa . Be 14-16 14-20 14-24 proportion of irides- PHT OIE HACK s oinciseis nies adie menide aor Quadrula pustulosa.............5. 14-16 14-20 14-20 cents. Tips are cut from all of the above-named shells. Take, for example, large washboard shells yielding 14-40 line blanks. The shells are first taken to the 40-line cutter, who cuts out all the 40-line blanks that are of the proper thickness with a true face (PI. XLIV). They are then taken to another cutter, who cuts out all the 24-line blanks that are avail- able. Finally, they are passed to the tipper, who cuts the remainder of the available material into 14 and 16 line tips. These tip blanks when run through the blank classifier may turn out a good per cent of blanks that are classed as butts, meaning by this that they are thick enough to make into any style of button; the tips are usually so thin that they must be finished with a machined face that requires the least material to be taken from it. At first glance the process of cutting might appear a very simple one, yet it is properly an operatiofi requiring much skill on the part of the laborer and the wisest type of management. A fuller discussion of the significance of the cutting room in the proper utilization of shells is given on pages 82-87. FINISHING PROCESSES. PREPARING THE BLANKS.—Before going to the finishing machines the blanks are usually passed through four intermediate processes. The blank classifier is essentially similar to the shell classifier on a smaller scale and need not be described in detail (Pl. XX XVIII, fig. 1); by falling between rollers the blanks are separated into different lots according to thicknesses (Pl. XX XVII, fig. 2). They are then placed in tumblers, consisting of heavy and slowly revolving barrels of iron or wood (PI. XXXV, fig. 2). In these the blanks are churned with water and pumice stone to clean them and remove the rough edges, making them easier to handle and more workable. Lye is sometimes used in connection with the pumice stone. Asa cheaper abrasive of more rapid action, fine sand may be used with the pumice stone. The blanks are now ready for the grinder, a machine fitted with an emery wheel which grinds away the horny backs and reduces the blanks to a uniform thickness (Pl. XXXV, fig. 5). These machines are operated by girls, who place the blanks face down upon moving belts 3 or 4 inches wide, while the belts convey the blanks underneath the emery wheels. These machines, as well as all others that require it, are generally connected by suction tubes with blowers for removing the dust that would otherwise But. U.S. B. F., 1917-18. PLATE XXXV. Tic. 1.—The churn used in polishing buttons. Fic. 2.—The tumbler employed for buffing blanks or buttons. (See (See p. 77.) Pp. 74.) Fic. 3.—A cutting machine of simple type. (See Fic. 4.—The shaker in which the buttons receive the final Pp. 71.) polish, (See p. 77. Fic. 5.—The belt grinder employed to remove the backs from blanks and bring them to desired thickness. (See p. 74.) Bu. U.S. B. F., 1917-18. PLATE XXXVI. Fic. 1.—The original foot-power lathe employed by Mr. Boepple in the inauguration of the fresh-water pearl-button industry. (See p. 65.) Fic. 2.—A modern automatic machine for shaping and drilling buttons. (See p. 75.) Buy. U.S. B. F., 1917-18. PLATE XXXVII. Fic. 1.—Tubular saws of different sizes, saw and spun fitted into the chuck, and shell from which blanks have been cut. (See p. 71.) Fic. 2.—Blanks of various sizes and thicknesses as cut from the shell and before submission to the “backing’’ process. (See p.74.) Fic. 3.—Finished buttons of several sizes and patterns. (See p. 76.) Buu. U. S. B. F., 1917-18. PLATE XXXVIII. The cutting room, where blanks are cut from the raw shells. (See p. 72.) Buty. U.S. B. F., 1917-18. PLATE XXXIX. Fic. 2,—The “‘finishing”’ room, where the blanks are converted into buttons by automatic facing and drilling machines (See p. 75.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 75 be obnoxious and injurious to employees (Pl. XX XIX, fig. 1). Grinders are paid from 15 to 25 cents per 100 gross, according to size and thickness of blanks, earning from $9 to $15 per week. Finally the blanks are again soaked in water to be softened for the finishing machine. In some cases, if too much mixed in quality or size, the blanks may be sorted by hand. MAKING THE ButTrons.—Having been classified, tumbled, backed, and soaked, the blanks are now ready for the essential processes of button making, which are accom- plished by an automatic machine of comparatively recent invention and of very ingenious design. The illustration (Pl. XXXVI, fig. 2) will aid in an understanding of the brief description of the working of the machine which can be here given. (Seealso Pl. XXXIX, fig. 2.) The blanks are fed by hand into depressions in the tops of vertical chucks, which are arranged in series constituting an endless chain. As the chucks in endless chain pass around the circumference of the machine each blank is automatically operated upon by various tools, and each tool is automatically sharpened and prepared for the succeeding blank. The processes accomplished in the machine consist in rounding the edges and carving out the center in the desired pattern to make the face of the button and in drilling two or fourholes according to pattern. After the first hole the drillrises, the button makes a turn through a fourth or a half of one revolution (according to whether it is to be a four-hole or two-hole button), when the drill again descends to make a new hole. After the last hole is drilled the chuck opens automatically to release the button, which is sucked into a tube connected with the blower system to be dropped into a bucket through a counting tube. Some twenty-odd distinct operations are combined in the double automatic machine, and it is interesting to record them. Let it be noted that the button travels in an oblong orbit, while the carving tools and the drills, respectively, travel in smaller circular orbits at opposite ends of the button orbit. 1. The traveling chuck, which is open after releasing a finished button, closes on the new blank placed in the top depression. 2. The chuck with the blank begins to revolve rapidly on its axis while continuing to travel to the right. 3. The face of the revolving and traveling button is applied to a carving tool of proper form to make the desired face. The tool itself is stationary on its axis, but travels in orbit with the buttons. 4. The facing completed, the tool rises. 5. The rotation of the blank is stopped. 6. The tool, continuing on its orbit, is sharpened on an emery wheel. 7. Before meeting another blank the tool is lowered by a small fraction of an inch to compensate for the shortening due to the grinding on the emery wheel. 8. The chuck, with its blank, leaves the orbit of the carving tool at a tangent to pass over to the orbit of the drilling tools. g. When the blank is in just the right position, one of the drills descends to make the first hole in the blank. In this operation the drill revolves, while the blank is station- ary on its axis, but both travel together. 10. The drill rises. 11. The chuck, with blank, turns through one-fourth of a revolution. 76 BULLETIN OF THE BUREAU OF FISHERIES. 12. The drill descends for the second hole. 13. The drill rises. 14. The blank turns another fourth of a revolution. 15. Third hole is drilled. 16. Drill rises. 17. Blank turns. 18. Fourth hole is drilled. 19. Drill rises. 20. Drill continues in its shorter circular orbit, to return into proper position for a later blank. 21. Button chuck rises a little and releases the button. 22. As the chuck passes beneath a suction tube the button is drawn up against a small, fine screen in the tube. 23. The button drops of its own weight upon a small trap. 24. When a number of buttons corresponding to a given weight have accumulated on the trap it releases and drops the buttons into a bucket. 25. The tripping of the door or trap registers the number of buttons finished. Another feature of the machine is the equipment of little screened suction tubes, some traveling, and some stationary, which draw away the dust whenever it is generated by carving or drilling. The amount which the carving tool may be lowered to compensate for grinding can be fixed by a large ratchet disk over the machine, which permits of adjustment to the one-thousandth of an inch. When the fisheye pattern (cf. Pl. XX XVII, fig. 3, buttons in second row from top) is desired, a thin, revolving emery wheel, or a steel fish eye cutter, is placed so that as the button passes from the carving orbit to the drilling (without central depression) the tool swings down and at one stroke cuts out the fisheye. There may also be an attachment for causing the blank to turn upside down, so that the back may be hollowed out instead of being left flat or rounded, as is ordinarily the case. This machine as described is the double automatic button machine. Somewhat older types are the single automatics, where separate machines embody the processes of facing and drilling. These are still in use in some factories. A very recent addition to the button machine consists in an automatic grinding and feeding attachment, whereby the blanks are first backed and then dropped into the chucks from an endless belt. As the machine is generally used, the blanks are placed individually in the chucks by the attendant, usually a woman, who becomes very expert. It will be recalled that the blanks were tumbled before being backed. In conse- quence, as they come to the automatic machine the back edges are slightly sharper than the edges corresponding to the inside surface of the shell. It is possible, therefore, for a deft operator to distinguish at the touch the outside surface from the inside and so to place the blanks in the chucks that the one side or the other (as desired) will be finished. There seems to be some difference of practice, nearly all manufacturers believing that the inside gives the best finish, while some find a better. product by finishing the outside. Experiments made at the Fairport station do not indicate a marked difference. Possibly a better or more uniform gloss is obtained on the outside, while a more pearly ‘‘water” results from finishing the surface corresponding to the interior of the shell. However, an obvious advantage in carving the inner surface Bui. U.S. B. F., 1917-18. PLATE XL. Fic. 1.—The churns, with buttons receiving polish by the use of acid and steam. Note the funnels from which acid solution drips into the churns. (See p. 77.) 2.—The sorting room, in which skillful operators classify buttons into «2 grades. (See p. 77.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 77 arises from the fact that the outside, after having been backed, is flat and true, while the inner face may have a natural unevenness. The blank on its back or flat side thus rests more securely in the chucks of the finishing machine and, when finished, both sides are in such true form as desired. Assuming each chuck to be properly filled in turn, a machine might finish from 33 to 72 buttons per minute, according to the speed given it. Ordinarily the capacity varies from 100 to 190 gross per day. In the days of handwork the output of a single operator in shaping was about 20 gross per day, and the buttons had then to be passed to the driller, who could turn out about 50 gross per day. The cost of a double auto- matic machine in r914 was about $1,300, with about $300 additional for the feeding attachment, but a machine would probably cost $2,500 in 1919. The operators are usually women, who were paid, in 1914, 1 to 2 cents per gross and earned $7 to $11 per week. In 1919 the rate of pay and earnings are 33% per cent higher. The automatic machine has revolutionized the industry of button manufacture from fresh-water shells. It makes possible not only a far greater yield, but a better uniformity of product than was ever possible with handwork. Something is yet to be desired in the way of lessening the amount of breakage of blanks in process of manu- facture, but the machine is being continually improved and perfected. PoLisHinc.—From the machine the buttons are taken to the churns, where they are tumbled, or churned, with water and pumice to clean them, take off the rough edges, and make them ready for receiving the final polish (Pl. XXXV, fig. 1 and Pl. XL, fig. 1). The polishing is also a tumbling process, in which, however, sulphuric or other acid is used in conjunction with steam. After the buttons are dried in shakers with sawdust (Pl. XXXV, fig. 4), they are placed with dry sawdust and washing powder in a combined tumbler and shaker. This process removes any trace of limy deposit and gives the final luster. Finally the buttons are conveyed in buckets or boxes to the sorting room. Sortinc.—A very important feature of a button factory is the sorting room, for the qualitiesand grades can not be sold if mixed indiscriminately. The classifying accord- ing to sizes and thicknesses has already been accomplished in the blank stage, but the grading according to freedom from defects of manufacture or from natural shell stains and with respect to color, luster, and iridescence must be accomplished by the hands and eyes of skilful sorters. Girls are always employed for this work on account of their superior deftness, or quickness of selection, and the most expert sorters can separate the buttons into 12 grades with extreme rapidity. They are provided with a well-lighted room and work seated in rows at long tables before windows (Pl. XL, fig. 2). The buttons are handled individually and thrown into series of boxes or drawers arranged about the operator; from 85 to 200 gross of buttons may be sorted in a day, so that sorters might earn, in 1914, from $5.25 to $12 per week, on the basis of pay at one cent a gross. In 1919, sorters are apparently earning from $10 to $19 per week on a sliding scale wage of 0.6 to 1.15 cents per short gross according to the number of grades (from 2 to 12) sorted. The number of grades varies with the several establishments, but it would not be practicable to enumerate them. Some factories make a specialty of iridescents or shiny-backs, as they were originally called. The iridescents are made from the hinder portion of the niggerhead, pimple-back, and related shells. If a niggerhead shell is 110307°—21——6 78 BULLETIN OF THE BUREAU OF FISHERIES. polished on the outside, there is seen to be an almost exact dividing line between the smaller brilliantly iridescent portion and the larger merely lustrous portion. Muckets and related shells produce no iridescent buttons. Some may be obtained from the blue- point, three-ridge, and washboard, but these are often not otherwise up to grade in quality. Buttons from these shells often require bleaching, and it has been observed that the process of bleaching increases the degree of iridescence. Although iridescents, when carefully selected, command a good premium, they are comparatively too few to make it generally worth while to work particularly for them. Care must be taken that the entire button is cut from the iridescent portion without overlapping of the forward portion of the shell. As a rule, no special effort is made to cut them, but as a number are cut incidentally, a premium may be paid to the sorters to separate them from the others. In consequence a limited number of clear iridescents are obtained which can be sold at a good price. When ordinary buttons of good grade were bringing 38 to 40 cents per gross, iridescents would bring 75 cents. Indeed, if there were any regularity of supply the price could undoubtedly be raised much higher and still the demand be good. BLEACHING AND DyrEInGc.—References have previously been made to the preva- lence of stained, spotted, or otherwise discolored shells. Such shells or portions of shells constitute a considerable proportion of the undesirable waste. Manufacturers have long striven to find proper processes of removing the discoloration without detriment to the quality of the product. Old methods of bleaching embraced the use of alkalies which injured the shell and caused the buttons to disintegrate or to break in the laun- dries. Hence bleaching came into disfavor in the trade, and some purchasers decline to purchase buttons believed to have been bleached. Bleaching as now generally practiced, however, is not injurious. Factories employing this process have each their own peculiar formulas or methods, but until very recently the essentials of the process were probably the same in most plants—peroxide, chloride of lime, and heat, with variations in the degree of heat and the period of action. Other chemicals are now employed in secret processes which seem to be very effective. Without bleaching, discolored buttons may be used to advantage for the produc- tion of smoked-pearl buttons, which are blackened by staining with sulphur and silver nitrate. Various dyes are also used in the production of fancy buttons of bright colors, as red, green, or blue, to suit the capricious demands of fashion. Some are now being so treated chemically as to produce an excellent imitation of the buttons made from the Trocha shells of Japan, but with better finish. Many shells in nature have beautiful colorings of purple, salmon, or pink, but the shades are not adapted to market demands, and it is claimed that the colors are liable to fade unevenly. Consequently, beautiful as some shells appear in natural condition, they must be classed as waste unless some effective process of bleaching or staining be applied. CARDING, PACKING, AND SALE.—Certain factories work for the bulk trade—that is, for the supply of garment manufacturers who do not require the goods carded; others for the carded trade exclusively, while some are prepared to supply both. If the bulk trade is supplied, it remains after sorting only to pack and ship; other- wise, the buttons must be attached to suitable cards. Sometimes the sewing is done in the factory; in other cases, more so in the past than at present, buttons are given out to women at home who wish to earn pin money at spare moments. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 79 Within the last two years a machine has been introduced for neatly attaching the buttons to the cards with small wires. It has the advantage of not only saving time, but of attaching the buttons individually, so that a single button may be removed without loosening the others from the card. This is a convenience to the consumer as well as to the manufacturer; for if, after carding, it is noted that a defective or an off-grade button appears on a card, it can easily be removed and another put in its place with a special machine. These wire-stitching machines are usually operated by girls, with one or two men to keep the machines in order. A single machine may turn out 150 gross per day, more or less. The record in May, 1914, was 240 gross stitched to the cards in 9.5 hours. The operators were then paid from seven-tenths of a cent to I cent per gross (about 1.2 cents in 1919), and might earn $6 to $10 per week ($8 to $11 in 1919). ‘The machines were not sold, but were operated on a basis of hire or lease. After carding, the buttons are ready to be boxed and sold. Individual manufac- turing establishments have specialties in the way of attractive cards and boxes or display outfits. Naturally the departments of advertising and selling have much to do with the success of the factory and constitute a considerable element of overhead expense. The wholesale price of buttons varies between wide extremes. Buttons of 45 lines and excellent quality may be sold as high as $2.25 per gross; or an overproduc- tion and accumulation of low-grade lines have led to sales at as low a figure as 1 cent per gross. Owing to the large number of low-grade buttons sold in bulk, the average price per gross, as computed from a statistical survey of the industry conducted by the Bureau for the year 1912, was about 23.5 cents per gross. As this price probably cov- ered freight, insurance, and discounts, the net factory price would have been corre- spondingly lower. The elements of cost in the manufacture of buttons must vary very widely accord- ing to the grade of the shell, the line of the button, the quality of labor, the size and output of the plant, and the systems of management. By courtesy of a manufacturer we may illustrate by the following citations of costs in particular lots of material which were carefully followed through the various processes of manufacture. The items in each case are as nearly correct as could then be made (1915), but the same tables would not apply exactly to other batches of the same lines. The data serve a useful purpose as illustrating relative costs of the several processes, although the prices of shells and the cost of labor have increased materially (approximately 75 per cent) since 1915. RELATIVE COSTS IN THE SEVERAL PROCESSES OF BUTTON MAKING, AS OF THE YEAR IQI5. Example Example | Example | of 28-line Items of cost. of r6-line | of 20-line | buttons, buttons, buttons. better . grade. Shell tn un it ac alas dSardtavtats = aVeis/e a RaR wena Aa dete es Dare/A Rae Coca catalase eee» $o. 0335 $o. 0524 $o. 1667 SERIA EMR ee saan Ce ia eirlsleiys Selelneywiciale ctevash ale's sisreldialeidda'y’sin.dix.a'ninyea(o >\slalowe singe» 209 +0523 «0689 1075 _ Overhead expense in cutting. .........-..eccpnnuer ener nercct eee ge eenceereneneeesnreees + 0196 + O194 +0393 Rear ierel Wietetetaiaikib atucaisracels: mpi nfo‘o'noiniS ataah testi wien ea alte ela sa‘a ele we cis eed neeba sabe nse = ye +0015 «0018 +0057 Overhead expense im grinding............-.0.00eceeeee eee eeec cece ee eeeee sects ren eenes +0125 +0125 +0125 ey ee tices, ctatatany wiaccaip aiatn elviaisl pia atest pee siamese oa: sim ales erormic ae /eipi= +O +O1 +Ors Overhead expense im machining.......5 0.06. cece ee ete ee ee cece tee nceeeneeneeeeenes +015 -OIs +015 SSOTGIN Goi ie ss wise sb Reine nicienietes oteirn aie nmaliin/se neta a bieie mnie bie at ne Se nisin ninaia wi .OL -Or - Or Ciyeetiend Sxpesise mi Sortie) i. iio. 252i le a Pea ee aE uke ee eee ee -014 ~ O14 - O14 _ Total (allowance for losses in process of manufacture not included).................. +1684 +2740 «3857 Een SE RRR Mot este oii aw pi viecoimy's Ae rs pin.w aie slag 'sime ce cece sia qen cgi siecle wg per cent... 20 26 43 80 BULLETIN OF THE BUREAU OF FISHERIES. Many of the cheaper buttons are manufactured, really, in the way of by-products; that is, the shells must be bought for the making of good buttons, and the cost of the shell is chargeable to these profitable lines; it is better then to cut the remainder of the shell into very cheap grades than to throw it away. ‘There are times, however, after the demand for the poorer grades has been oversupplied, that it becomes actually necessary to discard waste shell, unless one is to manufacture at a serious loss. UTILIZATION OF WASTE PRODUCTS. In 1912, according to a statistical survey conducted by the Bureau, 55,671 tons of shells were used for the manufacture of buttons. Assuming that only 90 per cent of this material became a waste product in course of manufacture, we find 50,000 tons of waste material. This waste consisted principally of shells discarded after cutting out the blanks, but a considerable quantity was in the form of finely pulverized shells or dust generated in the processes of cutting, backing (or grinding), facing (or carving), and drilling. This dust is not permitted to escape into the air, as otherwise the atmos- phere of a factory would be unendurable. It is removed and concentrated by streams of water played on the shell while cutting and by a system of blowers and suction tubes connected with the several elements of finishing machinery. The dust is, therefore, made available for use. The uses of the waste material would be various were it not for the fact that other cheaper materials are available for many of the purposes for which it isadapted. For instance, the waste shell might be burnt for lime or used for the improvement of soils or for many common purposes. The principal use of the waste shell is for the production of poultry grit, for which purpose it is prepared by passing it through crushing machines, which divide it into the desired fineness; defective or broken blanks and unmarketable buttons also pass into the crusher (PI. XLI). ‘The waste shells are sometimes used also as road-building mate- tial. The shells are very hard and do not pulverize so readily as oyster shell. In this respect there are obvious disadvantages as well as advantages. The dust is useful in stock food, and in condition powders for hogs and poultry; it serves also appropriate purposes as an element in the manufacutre of artificial marble, tile floorings, etc. It is said to form a constituent of some jewelry polishes, soaps, and cleansing powders. The present market for dust is, however, nearly negligible. In 1912, 22,530 tons of crushed shell were sold, yielding $114,722, besides about 10,500 tons of shell not crushed which were sold (probably largely to crushers) for $7,600. The sale of dust amounted to only 1,220 tons, bringing $3,470. The shell commanded about $5.50 per ton, and the dust about $4 per ton. In 1919 shell and dust yield, respectively, about $12.50 and $1 per ton. Both the shell and the dust possess certain exceptional qualities, and undoubtedly in time a better place in industrial uses will be found for them. USE OF SHELLS FOR NOVELTIES. The production of novelties from fresh-water mussel shells takes a wide variety of forms (PI. XLII). In 1912 there were six novelty works with an output valued at $61,800. From small whole shells or from portions of shells there were made such articles as watch charms, hatpins, stick pins, buckles, chains, cuff buttons, fancy buttons of all sizes and Buu. U.S. B. F., 1917-18. PLATE XLI. Crushing plant where waste shell and defective blanks are converted into chicken feed and other useful products. (See p. 80.) Buu. U.S. B. F., 1917-18. PLATE XLII. Novelties made from portions of fresh-water mussel shell. (See p. 80.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 81 many patterns, perforated tops for salt shakers, inlaid work, etc. Most or the novelty works are located in Muscatine, Iowa. The yellow sand-shell, being a rather long, straight-lined shell of peculiar pearliness, though not a clear white, is particularly adapted for novelty work, and thus commands such a high price that, under normal conditions, no manufacturer can afford to cut buttons from it. Shells of this species were often sorted out by the shellers on the river bank to be sold at a price two or three times the value of button shells, or from $40 to $60 per ton. Such shells as were not sorted out, but sold along with less valuable material, were subsequently again sorted out by the manufacturer for resale. Practically all of these shells were exported to Germany, where they were highly valued. Such shells for export sold as high as $90 per ton f. 0. b. New York. The Daily Consular and Trade Report of January 6, 1914, stated that there was a steady demand in Hamburg for the following grades of American shells, prevailing prices c. i. f. Hamburg per 220 pounds being quoted as follows: Equivalent price per ton of 2,240 pounds $5.95 $60. 58 4- 76-$13.09 | 48. 47-$133- 37 II.90- 16. 66 I2I. 16— 169. 63 Prices of marine shells at the same time were reported as ranging from $7.14 to $107.10 per 220 pounds (equivalent price per ton, $72.43 to $1,090.47). PROCESSES FOR POLISHING SHELLS.—Polished shells of the iridescent varieties make attractive souvenirs or table ornaments.. The polishing may be accomplished by one of the following methods, as described by J. B. Southall, shell expert of the Fisheries Biological Station at Fairport: Buffing process—The outer surface of the shell is ground off by an emery wheel or grindstone, the former being preferable, as it grinds much faster. If the surface of the shell is grooved, a file is generally used to remove the portions of the surface not touched by the grinding wheel. After the outer surface has been removed the shell is polished by holding it against a felt polishing wheel revolving at the rate of 2,000 to 3,000 revo- lutions per minute, fine polishing paste being applied to the surface of the wheel as needed. When all the emery scratches have been removed, the shell receives its final polish by holding it against a canton-flannel buffing wheel revolving at the same rate as the polishing wheel. If many shells are to be polished a double emery-wheel stand can be used to advantage by having the polishing wheel on one end of the spindle and the buffing wheel at the other end. Chemical process.—If the entire shell is to be polished—that is, the inner and outer faces—prepare the shell the same as for buffing. After the surface has been ground off the shells are placed in a cylindrical tumbler, using enough water to cover and a reason- able amount of fine pumice powder. It usually requires 8 to 10 hours of tumbling to remove the emery scratches and smooth the outer face of the shells. After the shells are smooth enough they are taken from the tumbler and placed in the polishing machine. Polishing machines can be purchased on the market, but a very good home- made machine can be constructed at a small outlay. Mount a short piece of I-inch shaft in a frame so that the shaft inclines at an angle of 45°, and at 82 BULLETIN OF THE BUREAU OF FISHERIES. the upper end fasten an earthenware jar of suitable capacity. The ordinary speed for the polisher is 40 to 60 revolutions per minute. The jar being tipped at an angle of 45°, the shells are tumbled over one another instead of resting on the bottom of the jar and moving with it, thereby allowing the acid to remain on the shells and cause pitting. A most convenient acid dropper is made by cutting a very narrow groove in the side of a cork and inserting it firmly in the tube of a glass funnel. With a little practice the number of drops to the minute can be regulated by the size of the groove in the cork. The next operation is to place the shells in the polishing machine and pour in a measured quantity of water just sufficient to cover them completely. A quantity of sulphuric acid equivalent to 20 minims for each 8 ounces of water in the jar is then placed in the dropper, which should be suspended over the polisher and so ad- justed that the acid will fall into the jar at the rate of 10 to 15 drops per minute. A very good plan is to take some of the water out of the jar and add it to the acid in the glass funnel, thus diluting the acid, and diminishing the danger of pitting the shells by allow- ing the pure acid to drop on them. It generally takes from 45 to 60 minutes for the polishing, if the acid has been gauged correctly. Just before the shells are polished the water becomes milky. Do not allow the shells to stay in the water long after the milki- ness appears, as the shells soon become coated with a white substance which is very hard to remove. After the desired polish is obtained dump the shells out and wash thor- oughly with clean, cold water; then wash the polisher, place the shells back into it, cov- ering them again with clean water, and revolve as before, applying steam to the water with a hose untilit boils. Just as the water comes to a boil pour in an amount of com- mercial muriatic acid equal to that of sulphuric acid used in the first operation and allow the shells to tumble a couple of minutes after the acid has been poured in. Remove the shells and wash as before. After washing the shells allow them to dry for 48 hours; then place them in a box tumbler and allow them to tumble in good, clean sawdust for a couple of hours. In this way the shells are buffed and receive the finishing polish. ECONOMY AND WASTE. THE PROBLEM OF CUTTING. In the early period of the button business, owing to the entire want of skill of the cutters and the apparent abundance of the shells, a most wasteful use of shell prevailed. Two or three blanks were cut from shells which should have yielded two or three times as many (Pl. XLIII). Factories could not now exist with such sacrifice of raw materials as then occurred. As a matter of fact, when new factories were continually forming in the early days, each eager to obtain cutters from other factories, it was impossible to main- tain system in the cutting department, and a confessedly unfortunate condition pre- vailed. With a condition of greater stringency bringing a necessity for better economy, the difficulties in the administration of the cutting department were augmented to the point of becoming a serious menace. Naturally, the need for betterment was first appre- ciated by the manufacturer; but the ways and means of bringing system and economy out of a condition of disorder.and waste were at first baffling. The process of cutting is at first glance a very simple one, and until rather recently the work of cutting was not generally understood to be skilled labor, nor was it really practiced as such. Nevertheless, in no part of the factory is the opportunity greater for A Buuv. U.S. B. F., 1917-18. PLATE XLIII. Illustrating wasteful cutting practiced in earliest stages of fresh-water pearl-button manufacturing industry. Compare ensuing plates. (See p. 82.) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 83 the exercise of care and good judgment on the part of the laborer or for intelligent and sympathetic cooperation between management and labor. The problem of the cutting room is, indeed, a complex one owing to the difficulty of maintaining an exactly definable standard or of arriving at a system of count and pay- ment that is at the same time neither open to criticism in some features nor conducive to waste. The desideratum of course, is a working plan, that neither denies to the cutter any portion of his reward, nor, on the other hand, relieves him of due responsi- bility for the cutting of shells in the best interest of manufacture and in accordance with proper usage of material. A great variety of shells are used, differing in size, weight, shape, and quality. No single shell is of uniform thickness throughout, but all taper, more or less, from thicker forward and center portions to thinner rim and tip; each shell, may, therefore, produce good, inferior, or worthless blank. By poor spacing an unnecessary portion of the shell may be wasted, or, from careless manipulation, an undue proportion of the blanks may be inferior or worthless. It is quite possible, too, to waste more time in the careful cut- ting of material than is warranted by the saving of material. Consequently it is essen- tial to impose checks in relation to the quantity and the quality of the output and to the proportion of product to materials consumed. DEFECTIVE BLANKS. Since no shell is of uniform thickness in all parts, the blanks from any given shell may vary from very thick to very thin. (See Pl. XXXVII, fig. 2.) All blanks wholly or in part thinner than two lines (one-twentieth of an inch) are called tips and can only be used for very inferior buttons.¢ A blank is never too thick to be acceptable, since it can be ground down to the desired thickness; unfortunately splitting is not yet practica- ble with fresh-water shells. A cutter, however, may cause a rim blank to split by twist- ing the shell when half sawed through, thus increasing the count though the resulting blanks are undesirable. Such an unfair practice is detected when blanks are found with- out a back, or covering, of horny epidermis. From too much haste or too little care, the blanks, instead of being sawed clean through, may be pushed out, leaving flanges of shell and horny matter on the outside which cause much trouble in the succeeding processes. In very thick shells it is often undesirable to cut the rim, on account of the blanks having such a pronounced bevel as to work poorly or to fly out of the chucks in the process of facing and drilling and perhaps injuring the chucks or drills. It is usually better, there- fore, to let this portion of the shell be wasted. Blanks cut through the eyespots (muscle scars) or through certain shell defects are sometimes undesirable. The cutter may space the blanks too closely, overlapping them, and thus producing buttons that are not round. In such a case, too, there is a danger of the saw being destroyed by the unrolling of the cylinder. (See Pl. XLV, lower left-hand corner.) This brief account of the more conspicuous possible defects in cutting will account for the practice of counting out certain blanks; that is, of requiring that a blank, to count for payment, must be two lines or more in thickness, must have the back on, be round, clean-cut, or without ragged edges, and be not cut through such spots or por- tions of the shells as may be prescribed. These specifications are simple and can be complied with by any conscientious cutter. The more difficult problem is that of getting the greatest number of blanks from the shells consumed. @ See footnote, p. 17. 84 BULLETIN OF THE BUREAU OF FISHERIES. CHOICE OF SHELLS FOR PARTICULAR LINES. Particular sorts or sizes of shell will work up most economically for particular sizes and grades of buttons. Accordingly, much discretion may be exercised in the apportion- ment of shells to the most appropriate uses, having in view the market demands to be supplied. A shell may be cut entirely into one size of blanks by one cutter, or the good blanks, the button blanks, may be cut out at one machine, while the remainder is passed directly, or after a lapse of time, to another cutter for taking out the tips. In other cases the shell is first worked for a few large blanks, and subsequently more completely used for smaller sizes. In this respect a large factory has a certain advantage over the small, independent cutting shop, since the latter must cut whatever shells are on hand into the particular lines for which the market is calling, while the larger factory may select from the bins the shells suited to the temporary needs, or may return to the bins, if desired, the partly cut shells. The detached cutting plant, on the other hand, may have an advantage in the saving of freight charges on useless shells. The assignment of the shells for particular lines depends upon the judgment of the foreman or manager alone; but a proper quantity and quality of output and the elimi- nation of unnecessary waste are contingent upon efficient labor in the cutting room. It will be of value in this connection to give some details of shop management, especially as observed in plants marked by the economical use of shells. SHOP MANAGEMENT. Disregarding minor differences, the shop management is, in brief, as follows: Each button cutter is assigned a machine, but is expected to furnish certain tools. The shells are weighed out to him, and he is assigned the proper line. Cutters with the best records may be favored with the most desirable lines. At certain intervals he turns in the blanks which he has cut and receives pay at so much per “count”. The count, or “gross, as it was formerly called, is not the ordinary gross of 12 dozen, but an arbi- trary and long-established unit of 14 dozen.? Since it would be impracticable to count the entire output of each cutter, his blanks, after shaking to remove chips and dust, are turned out into a tub on the scales and the gross weight recorded. A handful may then be taken at random to weigh out a given small unit in the pan of a balance scales. The buttons from the pans are then counted individually on the table, those not acceptable (for defects previously mentioned) being discarded from the count. The manner of count in different factories is divergent. In one, all blanks are counted as of the same value; in another, the blanks under four lines but over two lines in thickness are counted two for one. The latter plan is based on the fact that the thin blanks can be cut rapidly, and, although usable, are of relatively little value in manu- facture. The former plan, that of giving the same count to tips as to good blanks, is simply ‘‘an acceptance by the manufacturer of responsibility for the shape of the shell.” @ Thus we find that, in the language of button manufacture, neither the unit of measure, the line, nor the unit of count for blanks, the gross, corresponds to the common usage. The line is adopted from European practice and represents one-fortieth ofaninch. The gross of 168 was originally based on an allowance for breakage and other losses in the manufacture of the blank into buttons. The terms although unfortunate, are matters of custom and are understood by all immediately concerned. FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 85 Upon the basis of the sample count and the gross weight of the blanks the total count is readily computed for the purpose of payment. According to divergent methods payment may be made either at a flat rate per gross of 168 blanks, or the weight of shells used may first be charged against the sheller, and payment (at a higher unit rate, of course) is then made by deducting from the computed value of the blanks the cost of the shell at a fixed price previously agreed upon. The difference in the two methods has to do with the fixing of responsibility for the economic use of shells. In the first case, the management alone is responsible for economy, in the second the responsibility is divided. The latter seems really fairer to the cutter than the former and puts a proper premium upon careful and efficient work. In case of payment exclusively upon the basis of blanks returned, the interest of cutter and management are continually in conflict. The cutter is tempted to sacrifice economy of material to gross output of blanks per day, while the management must be continually endeavoring to prevent undue waste of shell. This is a condition provocative of friction and liable to result in discharges or in docking. In the other case, the cutter understands that waste of shell diminishes his earnings. Suggestions from the foreman to show how more blanks could be taken from a given shell are, therefore, received as personal help rather than as rebukes. A single instance that came under the observation of the writer will serve to illustrate: A cutter had laid down a cut shell when the foreman pointed out how two more blanks might be cut. The cutter readily accepted the suggestion, which was to his own interest, and as the manager went away remarked to the writer: ‘“‘Under the old system there would have been a regular ‘call down’ about that.” Cutter and management soon learn that care and system profit more than haste. The cutters earning the best wage are those who begin at the right place, plan out the cutting to use the most of the shell, cut in rows, and take the time necessary to avoid mistakes. (Cf. Pls. XLIV, XLV, and XLVI.) Under such a plan there can be no occasion for discharge. The cutter who lacks the intelligence, the aptitude, or the character to become a skillful cutter, even with help, must find the business unprofitable and seek employment in other lines for which he may be better adapted. The work passes naturally into the class of skilled labor, and the skillful do not have their proper earning diminished by an average rate which is lowered by the waste and scant production of the unskillful. This report is not, however, concerned with the relative merits of systems of man- agement except as they may affect the waste of raw materials. Propagation, protec- tion, and the economical use of material are phases of the general conservation problem that are indissolubly linked. Without regard to the details of any system of adminis- tration, it may be said that the economic use of shells will necessarily be promoted by a plan (1) which divides between management and labor the responsibility for waste of shell; (2) which does this in such a way as to remove as far as possible the necessity for arbitrary docking or rebuke or discharge; and (3) which consequently substitutes for such forms of discipline a true spirit of cooperation between employer and employee for mutual advantage. The best spirit and the best intelligence of all concerned may well combine for the most advantageous use of materials, having in view both the dimi- nution of waste and the improvement of quality of product. It is futile to imagine 86 BULLETIN OF THE BUREAU OF FISHERIES. that in the long run the interest of cutter or manufacturer is promoted by waste or antagonism to legitimate improvement. PROPORTIONS OF PRODUCT AND WASTE. The unavoidable waste in the commercial use of mussel shells is remarkably high, assuming the most economical use possible of materials under present conditions. The waste involved in the combustion of coal is often cited as an example of unavoidable loss, where, under the most efficient methods in use, only a small per cent of the latent energy is converted into power. In button manufacture we find that only 5 to 8 per cent of the original gross weight of the mussel enters into the button product; but the re- mainder in this case is not all lost, since there are waste products which are utilizable at a less profit. In the first place, when the mussel is taken from the river, we find that about 3 per cent of the dry weight is thrown out as meat.? The losses in the shell at different stages of manufacture, as determined by averages from several specific tests made by J. B. Southall, are shown in the following table: LossES IN SHELLS OF CERTAIN SPECIES DURING MANUFACTURE OF BUTTONS. Lake . Waste or by-product. Pepin a Per cent. Diseardeéd dhielf. ct cct SATA PETES. FAN, OT. TSS TT Abs Si awison c ctheily oe add. smn eater eee «8 60. 8 73-6 Dust in sawing blank «1... 1... 5-2 eee e ee eee ener een eet teen teeter neta e ene ee tees eeteneee ee ee ses 16.9 8.8 Dust in grinding and finishing button ............-... PMG cincate ROME Se a wtdtate cat URAL WS 12.9 10.8 Total wasteor By-Products 0)5. cjeicid Soe o Liaewlals me aie kidiciem slablolaleiels vleldslem a WaN Renee b RAT ae sv eth e lctetnls WAM 90. 6 93-2 Weight of Hitt tons os. oly sseln s aes cians Desai Wee erat ae aera ae ere one ie elas einai ai 9-4 6.8 Lh» t BEEP e ROE AR cet dtc AGREE” DRECOR SEE SEOs IAneE nema ESEene anne Crnpeeoarrnormcrer code Sande 100. 0 100.0 Roughly speaking, 7 per cent of the total weight of heavy shells like the niggerhead is marketed from the factory in a form worth $2.16 per pound,® while of the remaining 93 per cent, a portion is entirely thrown away and another portion sold as crushed shell, or dust, at a quarter of a cent per pound.° The table and data are not of purely academic interest. They point to the signifi- cance of the problem of the utilization of the now unavoidably wasted material, and they emphasize the importance of putting more of the shell into the high-priced product, the buttons. It remains to differentiate the instances of waste which are prevented by correct practice in cutting and those which arise from the form or character of the shell, and which consequently may be obviated only by new discoveries in method or by changing demands of the trade. WASTE IN CUTTING. It may be conceded that there is some waste which it would be possible but not desirable to avoid. Given the present economic conditions, it will appear that parts of the shell which could be cut are better left uncut, because the cost in labor @ The proportion of dry meat to shell varies widely with the different species. It is safe to say, however, that, on the average, the meats with all water dried out represent one thirty-fifth of the total dry weight of the mussel. The utilization of the meats is discussed in part 2, page 6r on the mussel fishery. > Figuring 1,248 buttons to the pound at 25 cents the gross (good quality 16 line). ¢ Price of crushed shell, $5.50 per ton. Price of dust, $4 per ton. Buu. U.S. B. F., 1917-18. Fic. 1.—Poor cutting due to want of care or lack of experience. Note poor spacing, overlapping of blanks, and exposed margins of shell, indicating that blanks were pushed out. (See p. 87.) Fic. 2.—Illustrating careful, methodical cutting, (See p. 87.) Buu. U. S. B. F., 1917-18. PLATE XLV. Ee & | Illustrating reasonably careful practice in cutting. In endeavoring to cut too closely the shell in lower right-hand corner the saw was destroyed. (See p. 87.) But. U. S. B. F., 1917-18. Pyate XLVI. Fic. 1.—Illustrating economical use of shells of various sizes. (See p. 87.) Illustrating theoretically correct use of shell for small blanks. (See p. 87,) FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 87 or in injury to machinery would more than counterbalance the saving of material. Nevertheless, all factories are not equally efficient, and some waste occurs that could be prevented by watchful and judicious management and by proper training of labor. It will not be out of place to illustrate this fact by the use of a few photographs with brief explanation. Plates XLIV, XLV, and XLVI illustrate the progress in good cutting. Plate XLIV, figure 1, shows a single shell where the cutter, though exerting some care to avoid waste of material, showed a lack of experience or judgment. The blanks were irregularly and poorly spaced. In some cases, too, the blanks were spaced too closely, so that true circular buttons could not have been obtained. Contrast this with the lower figure on the same plate, where the cutting was in rows, with best practicable use of shell. Plate XLV includes shells from various plants where the cutting is done with more than average economy; the shell in the lower right-hand corner illustrates a danger in too close cutting, where, because of overlapping a space previously cut, the cylin- drical saw was caused to unroll with more damage than a slight loss of material would have entailed. The shell was evidently thrown away in disgust. The avoidance of the rim blanks in the heavier shells of this illustration may be noted. Blanks obtained from the rim of such shells have so much bevel that they will not work successfully in the further processes and are liable to cause injury to machinery. Plate XLIV, figure 2, and Plate XLVI illustrate the most economical cutting feasible under present conditions, and also show how shells are used to practical advantage by double cutting; that is, by taking out the larger lines first and then removing the smaller lines. Plate XLVI, figure 2, shows an excellent spacing of blanks in a very favorable shell. WASTE DUE TO EXCESSIVE THICKNESS. A serious source of waste is found in excessive thickness of shell. If a shell is of sufficient thickness for buttons throughout, and in no part of excessive thickness, the waste consists in the spaces between blanks and the unutilizable portions of shell such as the hinge, umbones (knuckles), and rim. Many otherwise excellent shells, however, are very thin at the tips, while the blanks from the forward portion of the shell are so heavy that from one-half to three-fourths or more of the thickness must be ground off (Pl. XXXVII, fig. 2). It is unfortunate that such blanks can not be split. Very thick blanks from ocean-pearl shells may readily be sliced or split into as many blanks as desired; but in the fresh-water mussel shells there are irregularities of stratification, or faults, in the shell which cause the splitting to occur in such irregular fashion that with any method or device so far employed the waste is quite out of proportion to the saving. Between shells which produce 700 gross of buttons per ton and those which produce 1,000 gross per ton there is a distinct difference in economic value, assuming quality to be the same; also, there is a saving in the working of lighter shells, since there is evidently a useless waste of time involved in the sawing of thick shells and in the subsequent grinding away of excessive thickness. Furthermore, the heavy shells are of much slower growth. Assuming an equal quality from shells which are lighter and more nearly uniform in 88 BULLETIN OF THE BUREAU OF FISHERIES. thickness, there will be a quicker return from efforts at propagation of such species and a greater likelihood of being able to maintain the supply. There will undoubtedly develop a more insistent demand for the best. yielding material, causing an advancing price, while the shells which are found to work with less economy will decline (relatively) in price to-a point where they can be used with actual profit. WASTE IN DISCARDED SHELLS. A discussion of the general subject of economy would not be complete without reference to the discarding of certain classes of shells which have some good qualities, but which for one reason or another are not suitable. In the very beginning it was thought practicable to use only shells of comparatively uniform thickness. Accordingly, muckets and sand-shells were bought, while niggerheads, pimple-backs, etc., were refused. It was not long before the excellent qualities of these shells became apparent, and it was found practicable to cut them. The discards at the present time are almost exclusively shells which could not yield buttons for which there is a market. Such are thin shells, colored shells, and shells exceedingly stained and spotted; notallof these are wasted. Some pink and purple shells, when thick enough, are found to be of a particularly good working quality; they can be used for making the smoked-pearl buttons by staining with silver nitrate. The demand for such buttons is limited, and, as many pink or discolored buttons are cut incidentally and culled from the better grades, there is little actual market for the pink shells beyond the limited requirements of the novelty trade. The elephant’s ear.is a rather common shell of beautiful pink, purple, or salmon color, and it is said to be superior in working qualities to the better grades of white shells. It is unfortunate that a simple, satisfactory method of bleaching has not been available or that there is no market for the natural pink and purple buttons produced from it. Many discolored, shells may be bleached, though somewhat imperfectly, but bleaching methods have been greatly improved in recent years. The discoloration is generally attributable to disease or parasites. Discolored shells seem to be more common in sluggish water and in portions of streams polluted with * sewage. : RESUME OF MANUFACTURE. The fresh-water shells are used preeminently in domestic button mantfacture, though a small proportion enters into the producton of novelties, and up to 1914 an increasing number were being exported both for novelty and button making. The process of button manufacture consists in classifying and soaking or moistening shells, cutting (either in detached cutting plants or in a room of the complete factory), cleaning, tumbling, backing, and soaking the blanks, facing and drilling (in one or two automatic machines), cleaning, polishing, drying and sorting the buttons, and finally sewing them upon the cards or packing in bulk. Processes of bleaching and of staining may be introduced as desired. The principal operations of skill are the cutting of blanks and the sorting of buttons, The chief desiderata are the perfection of an automatic cutting machine and the elimina- tion of waste as far as practicable at all stages. These needs are receiving the careful FRESH-WATER MUSSELS AND MUSSEL INDUSTRIES. 89 attention of many manufacturers and mechanics. Remarkable improvements in button- making machinery and systems of management have taken place in recent years. Perhaps conspicuous advances in the future may be made, not only in such improvements of management and existing machinery as are always to be expected, but in the inven- tion of an efficient machine for blank splitting and in the perfection of bleaching processes; also in the greater utilization of unavoidable waste. In 1912 there were 196 separate plants employing mussel shells in manufacture. Of this number 153 plants were devoted to cutting only, while 36 factories engaged in finishing and grading. Of these latter 20 included cutting rooms, also, and thus comprised all the processes of manufacture. In addition there was a single branch plant devoted exclusively to the grading of buttons. There were 34 shell-crushing plants, of which 32 were connected with button factories, and there were 6 novelty works. ‘These establish- ments were located in 20 States as follows: Alabama, Arkansas, Kansas, Kentucky, Illinois, Indiana, Iowa, Maryland, Massachusetts, Michigan, Minnesota, Missouri, New Jersey, New York, Ohio, Oklahoma, Pennsylvania, Tennessee, West Virginia, and Wisconsin. The industry is peculiarly American. The material has until recently been obtained in no other country, and the machinery and methods are largely of American design and development. TAT AONE scadiens Sata anal rters otg hd inf Dikas Satta! Wie clita nk arpa hy 25 yet) ete ey . Ve Lid shivonny w Bit LIFE HISTORY OF THE BLUE CRAB w By E. P. Churchill, Jr. Assistant, U. S. Bureau of Fisheries SAR). add BHT ¥O YAOT2 FI ie ; Se 2 “hg Bee Al Sidon) 4 A : by! Yorvomes A to anal, 27. AS Areoteya? CONTENTS. & Page RU EAnET es Oe are ate etn serene era er asietolis/aVeNr ic) = etceysie muatane ate tniierojngrsmamnnser sivicunte ere. siciais #)sisleseieie)s/eieie'sis, /eys\sheelaists 95 Testa eenal GEE EERO: oqo pecose aces Abobo: Uda espe dam loo ocd CONSE CDCOn OOOO COCO GUO Sc. 95 bepelapaiene of THe VOUS 6-600 5.0 ieee nea cee Ste tne wma eet wieder aereinineescen see tenes 96 NE a teen Sercle el chre dais: rind as rig ustenisie Siem t hrelgad yp rlecicteige sia ee eninele se 104 tea ita? cso npaceeanbe te 0e60be dae Bo g6Gtr bo UD DUS na pape AGU otOOs aera LOCH BOO ULES ET OG 106 Migration in Chesapeake Bay.......-..--2-20sseeeeeeerenesecseeeee sree receesesceerseeeec ces 108 shail GE iGeos see mandoc vo aupe nb ddoboo dn seosp mob oopmopNGuodocrar OBC OIGOUROODROOROI oS 113 Pipe eresapte WN es 37 elcol olsicle eran atmesnrnie wie rincin'a s siaiv rit Mn siePeineis notin) icieeleintainciele ri II5 ee Rte eae ein ane ecn vice nina Vrnciae rire vide «PAS seen eccmiee ences II5 Spawning experiments with crabs taken during the winter. .........-.. 202s sess erences 119 Number of batches of eggs laid. ............ 0.0 cece cece eet ence nee e neste neste settee cen e es 120 Wiaviivac Title iwioa pee bon obaobb pan pba ace qube cond miaapbo s gues EapopoUdaers no OSU aU COCR CRIES Tt 123 Length of Ia eee pects cay ies a aah oh ots caxe hn mia oy ea item eyed ee, = wuss cas sal eel oem wyie\scs: #)*'-miebe i eles Ng 124 Gi Ach dana uiked: eds Don Be ae GnDUS EMA RubcRoaciuneiads Garni GE eon ctoOG Gor EaeoH imma Soo or: 125 Reerleniapianiak plated... 2p os) srin amie cin arse ee ono rivet mismieimenes emlunielees weiss s ncipeteee 126 110307°—21——7 93 ee mm tbe dS Pusey a > SISSERE “pont Sa aaeent? fi cette a _.. tivtedt fereareey oD. atinieeang ra oad ae ve ehasmieywe gelmerege bist s £ os eer ted 7 i sada VE "aid ty ge Cie oe #sieicy lo o othe it Tm LE TOEOEE. iBUEE We Ss ba be, 1LOl7Z—1S: PLATE XLVEL. Coloration of the Crab in Molting Stages. BREUKER & KESSLER CO, PHILA. LIFE HISTORY OF THE BLUE CRAB. x By E. P. CHURCHILL, Jr., Assistant, United States Bureau of Fisheries. em NAME. The crab commonly used as an article of food on the eastern coast of the United States belongs to the species known as Callinectes sapidus.* ‘This name may be freely translated as ‘‘savory graceful swimmer.’’ ‘Those who have eaten this crab in either its hard or its soft-shelled state can give ample testimony that the term ‘‘savory”’ is well applied. It is known generally as the blue crab from the fact that considerable blue color is usually found on its upper surface, especially on its claws. In Chesapeake Bay regions, where it forms the basis of an extensive crabbing industry, it is referred to simply as the ‘‘crab,”’ other species of crabs having some common distinctive name, as the ‘‘fiddler,’”’ ‘‘sea spider,” etc. The blue crab belongs to the family known as Portunide, or swimming crabs, from the fact that the posterior pair of legs, or back fins, are flattened for use in swim- ming. All crabs of this family, in which the abdomen, or apron, of the male is J- shaped, belong to the genus Callinectes. The species Callinectes sapidus includes only the edible, or blue, crab of the Atlantic coast of the Americas. HABITAT AND DISTRIBUTION. The blue crab is found on the Atlantic coast from Massachusetts Bay to at least as far south as the northern part of South America. In the United States it is common from Massachusetts to the southern extremity of Texas.? Although occurring at most points on that part of the coast, it is especially abun- dant in the bays and mouths of the rivers. It is found during the summer in relatively shallow water but at greater depths during the winter. Although its natural medium is salt water, instances are known in which specimens have been found in brackish,? and even in fresh, water.°® The adult crabs tend to remain in deep water, but the young, especially, come inshore to a point where the water is only a few inches in depth. In general, the closer inshore the observations are made, the smaller the size of the crabs found. @ This name was established by Dr. Mary J. Rathbun in The Genus Callinectes. Proceedings, U.S. National Museum, vol. 18, 1895. Washington, 1896. > Hay, W. P.: The Life History of the Blue Crab (Callinectes sapidus). Report of the U.S. Bureauof Fisheries, 1904, p. 400. Washington, 190s. ¢ A crab dealer of Hampton, Va., related to the author that he found numerous crabs in Back Bay, Va., on Nov. ts, 1917, the water there being fresh enough to drink. 95 96 BULLETIN OF THE BUREAU OF FISHERIES. The blue crab is especially abundant in Chesapeake Bay. This body of water is of sufficient size to afford a breeding ground for an immense number of crabs. ‘The young are very abundant in the region extending from the vicinity of Tangier Island, Va., to Baltimore, Md., the bottoms underlying the shallower waters of this part of the Bay forming, during the summer, a ground especially suited to the growth and molting of the maturing crabs. In the deeper waters of the southern part of the Bay the adult crabs lie on the bottom in vast numbers throughout the winter months. During the summer they frequent the more shallow waters, where they spawn in great abundance. DEVELOPMENT OF THE YOUNG. The young of the crab are hatched from eggs (Pl. XLVIII, figs. 1 and 7, and Pl. LIT, fig. 30), which measureabout 1/100 of an inch in diameter, * not as large as the period at the close of this sentence. When first laid, the eggs are yellow or orange in color, due to the color of the yolk granules within them, which serve as food for the young as development proceeds. As the eggs near hatching, this color disappears, and, since the eyes of the young are comparatively large and are of a very dark color, the mass of the eggs appears almost black. As the eggs are extruded from the body of the female they become attached to the fine hairs of the swimmerets on the under side of the abdomen. There are no swim- merets on the anterior segment of the abdomen, but there is a pair each on the second, third, fourth, and fifth segments, and none on the sixth or seventh. ‘There are thus eight swimmerets in all, four in a row on each side. Each swimmeret is made up of two branches, an inner and an outer (Pl. XLIX, fig. 16). The hairs borne by the inner are much finer and longer than those on the outer, measuring in diameter from 1/200 of an inch at the base to 1/700 at the middle.” The eggs all find lodgment on and are carried entirely by these hairs of the inner branches of the swimmerets, but never by the outer. Microscopic examination of these hairs, when they are bearing eggs, shows that each hair is covered throughout almost its entire length with a thin coating of a semitrans- parent material of a faintly yellow color. Each egg is attached to this covering by a separate short tendril of the same material, the hair and its burden resembling a long thin stem, with a great number of berries attached to it by short tendrils (Pl. XLVIII, fig. 7, and Pl. LII, fig. 30). From this resemblance, the crab, when bearing eggs, is sometimes said to be “in berry” or “‘berried.” There are eight tufts, or clumps, of eggs, corresponding to the eight inner branches of the swimmerets. These tufts are so large, however, that they are all crowded together, so that there is formed a flattish mass about 3 inches wide by 2 long by 134 deep, and fairly smooth in contour (Pl. L, fig. 19, and Pl. LI, fig. 22). From its general appear- ance and color, this mass is known commonly as a sponge, orange, lemon, punk, or ball. The egg color is yellow or orange when first laid, but, as already stated, it becomes almost black as hatching time approaches. The abdomen is pushed back by the sponge until it extends almost in a straight line with the body, except at the posterior end, where it curls downward behind the mass of eggs. a ‘The measurement given in this text was made by the author. The size is placed at 1/108 of an inch by F. H. Herrick in Natural History of the American Lobster. Bulletin, U. S. Bureau of Fisheries, Vol. X XTX, 1909, p. 310. Washington, rorr. b Herrick’s op. cit., p. 310. LIFE HISTORY OF THE BLUE CRAB. 97 The number of eggs contained in a sponge of average size is enormous. S. I. Smith places it at 4,500,000,4 which number was quoted by Herrick.’ Paulmier ¢ estimates that between 2,000,000 and 3,000,000 eggs are borne in the sponge. The present author found that by actual count there are about 200 eggs upon each hair of the swimmeret. The hairs occur in fairly regular bundles of about 5 each, there being about 20 bundles in each of the 11 rows, arranged in a longitudinal manner, on the swimmerets. There are 8 swimmerets. Computing these figures gives the sum of 1,760,000. It must be borne in mind that this figure is not much more than an estimate, as it is next to impossible to determine accurately the number contained in such a large mass of objects as minute as the eggs in question. ‘The most accurate statement that can be made is that there are from 1,750,000 to 2,000,000 eggs in a sponge of the usual size. The eggs had been fertilized while in the body of the female. This process is described on page 117. The eggs are carried upon the swimmerets while their development goes on, or during what might be termed the “‘period of incubation.” About 15 days are required for the eggs to hatch. A female crab was kept under observation in a float (Pl. LI, fig. 20, and Pl. LV, fig. 37).2 On June 15 this individual threw out a sponge. On June 29 it was found that some of the eggs had hatched, since there were many empty shells upon the swimmerets. By July 2 nearly all the young had hatched out and left the mother. In this case it will beeen that the period of incubation was from 14 to 17 days. Another crab was observed to spawn on August 15. The eggs hatched within 12 to 15 days. The temperature of the water, no doubt, has some effect upon the duration of the incubation period. During the last of June, when the first experi- ment was being carried on, the temperature of the water was about 79° F. During the August experiment the water was about 85° F. This may account for the fact that the eggs hatched somewhat more quickly in the latter than in the former case. It has been thought by some that the young crabs cling to the swimmerets of the mother for a time after hatching. Binford,’ however, observed the young as they hatched from two females and found that this was not the case. The present author, in the case of the two crabs used in the experiments just described, found that the shell of the egg split into two parts (Pl. XLVIII, fig. 5), the young crab emerged and, after freeing itself from a thin membrane which covered it, swam away. Numerous empty split shells (Pl. XLVIII, fig. 10, and Pl. LII, fig. 30) were found on the swimmerets of the adult, but no young crabs were observed clinging there. Several other crabs were observed as the eggs were hatching, but in no case were any young found clinging to the swimmerets. It is thought by many that the young, immediately upon hatching, turn about and devour the mother crab. Needless to say, this idea is a mistaken one, although, of course, quite small crabs feed upon and may even consume any dead crab which they chance to find. In fact, this erroneous notion arose from the occasional observation of @Smith, S.I.: Report on the Decapod Crustacea of the Albatross Dredgings. Report of Commissioner of the Fishand Fisheries for 1885, pp. 618-619. Washington. =) » Herrick; op. cit., p. 309. ¢ Paulmier, F.C.: The Edible Crab. ssth Annual Report, N. Y. State Museum, rgor, p. 134. d Unless otherwise stated, all of the experiments discussed in this paper were carried on at Hampton, Va., between October, 1916, and October, ro17. € Binford, R.: Notes on the Life History of Callinectes sapidus. Johns Hopkins University Press, February, ror1, p. 1. 98 BULLETIN OF THE BUREAU OF FISHERIES. a dead sponge-bearing crab being surrounded and devoured by a multitude of young crabs about the size of the fingernail and the assumption that these young had just hatched from her sponge and then turned about and were devouring her. From the description, given in the following paragraph, of the young immediately after hatching, it will be seen that it would be impossible for them to devour a hard-shelled adult crab, even though observation were lacking to disprove the notion. The young of the blue crab, after hatching, pass through two stages before assuming the true crab shape.?_ In the first stage a young crab is known as a zoéa. The zoéa is virtually microscopic in size, measuring about 1/25 of an inch in length. From Plate LI, figure 21, it will be seen that in this stage the crab is much unlike the adult form. The body is somewhat cylindrical in shape, the eyes large and conspicuous, the spines at the sides short; there is a long curved spine on the back; the claws are lacking; and the abdomen is long and round, ending in a sort of forked tail. The zoéa has a long, sharp beak, two pairs of antennz, and four pairs of leglike appendages. The true legs have not yet appeared. The zoéa swims backward by very rapidly jerking the abdo- men up against the lower side of the body. The crab in this stage is free-swimming and does not crawl over the bottom, as it does in the later stages. The zoéa increases in size only when it molts. At the present time, however, it is not known how many moltings occur before the second stage is reached.? At each molting the new form resembles a little more closely the next stage. The crab in the second stage is known as a megalops (PI. LI, fig. 23). It is still very small, being less than 1/25 of an inch in width. The megalops more nearly resembles the adult, having a rather flattened body and an abdomen shorter and wider than that of the zoéa. ‘The eyes, however, are as yet more prominent than in the adult, and the two posterior legs are not flat, but rounded, and each is provided with a sharp point. The abdomen is not curled against the under side of the body, as is the case in the adult. The megalops swims freely and, also, may walk on the bottom. It is as yet unknown how many times the megalops molts before taking on the true crab shape. Smith and Hyman (op. cit.) found that, in the case of the rock, the green, and the fiddler crabs, there is only one megalops stage, the first megalops molting directly into the first crab stage. It is quite probable that this is true also of the blue crab. Whether or not this is the case, there does come a molting at which the megalops suddenly assumes a shape very similar to that of the adult, except that the width of the body is not much greater than the length, and the eyes are borne on larger and thicker stalks. This creature may be called the first crab. The crab, as well as the preceding forms, increases in size only at the time of molting. Up to the present time no one has observed a crab as it passes through all the different molts involved in its life history. Certain stages, however, of the lives of several different crabs have been observed, so that a fairly accurate estimate can be made of the number of moltings which occur and the time required for the crab to reach the adult stage. a The description of the first two stages of the young crab is abridged from the Handbook of Invertebrate Zoology, W. K. Brooks, published by S. E. Cassino, Boston, 1882. > Paulmier, op. cit., p. 135, estimates the number of moltings to be probably six, but gives no data upon which to support his claim. Smith, S.,in The Invertebrate Fauna of Vineyard Sound, U. S. Fish Commissioners’ Report, 1873, found that there are four zoéal stages in the green crab of the Atlantic coast. Hyman, O. W., in a yet unpublished paper found that the fiddler crab passes through five zoéal stages. LIFE HISTORY OF THE BLUE CRAB. 99 Dr. Binford observed a crab which was kept in an aquarium while it passed from the megalops stage to that of the sixth crab stage.* The following table sets forth the results: TABLE A. Date molted. Stage. bi al THerease: (eee a| [SS Inch. Inch Jaly 18.5.5 6 ets soe eee ce AM eae ae naloeh onan seaisitoweusendaeesees Megdlopsi i... csc. ccccteceweesnce aT, ol smartest JULY 29. .06-- eee occa e eee ce eeec ne tene esos aaaesscnedeseereesedecnsses Rirepirabyeee..civecnies sides samiciss 128 0. 088 Jully 27.0... 2.00 e nce e eee e en nnee nec enensinwenesscesseceaeenenccnsoeess Sore CSA sein seas seis easels 196 068 ATIGUSt in sei oo occ on Shlania enw owing ules geen alee MMe dene ee gue n sane dij hint ae ls ya e ANE So Seo Goa ee 260 064 August 6.....55 0-00 ec cece dense essen eee bane csen ced ddedecsocctsrerce Dini tid les |? SSG Ee Ip BOOOU a oer ic 348 088 Avig“ist 12.....5-2 0c 0ccccecednccec cee ereberterssceeddendeceeedeceaees ANGER CTA D Gace ascii eotsw a bins oats 456 108 PASIGUISE AS. ski soe eee smiewidtne ccs cen wale one anes baiele diese se sscnee SEMA eedtatine age aopronese +516 060 This crab was caught in its natural habitat while in the megalops stage and placed in the aquarium on July 18. It changed to the first crab on the next day. The zoéa and megalops stages, therefore, were completed by about the middle of July. In the region of Beaufort, N. C., where this experiment was carried out, the bulk of the young hatch during June; therefore, probably not more than one month had elapsed between the hatching of this specimen and the time of confinement in the aquarium. That is to say, not more than a month was required in which to com- plete the zoéa and the megalops stages. The author succeeded in carrying several crabs through certain of the molting stages between the sixth crab and the adult. Some of the crabs were kept in floats, but most of them were confined in cages of quarter-inch mesh wire (Pl. LV, fig. 38). These were placed on the bottom, in the water of Hampton River, at a depth of 4 or 5 feet at low tide. Each cage was equipped with a strand of wire by which it was lowered and raised and which was attached at its upper end to a stake. Although it was impossible to carry a particular crab through its entire life cycle, by begin- ning the experiment with crabs of different sizes, the author was able to collect data, from which it is possible to estimate how many moltings occur as the crab develops from the sixth crab stage to the adult size. It must be kept in mind that at this time of life the successive stages of the crab are very similar, except in size. This last also varies with the individual, the temperature, etc., so that it can not be ascertained by examination in just which stage a particular crab is. The best that can be done is to try to form as accurate an estimate as possible of the number of times a crab usually molts while reaching adult size and the time required to attain maturity. The results obtained by the author are presented in Table B. @ Binford, op. cit., p. 2. 100 BULLETIN OF THE BUREAU OF FISHERIES. TABLE B. Width. wate Time be- Individual and sex. at hh lo Increase. tween Before. After. molts. Inches. Inches. Inches. Per cent. Days. INO, (5p 20ers a sterajeia nee oO sia hie deine Sia races Sein ards ereieiat sig Solel iae ©. 437 0. 562 0.125 28 1) ‘62 812 a Napa) See eee este vente aaa tala hinn seen eel eee ae aan een ale te 2 a ag () -812 1-000 +187 23 13 +750 (2) IRE aasseceecleee nee ebee (2) Noits, female. Fi... os:c<.o pies oelps Varoleelarerasran et wee eee a aa sale Gi (>) TAGO' | as haoe wed sews ee 1r 1.000 I.250 +250 25 10 IND: 4; Lette eee ccnciccs ciistsic cis'eesiestaisesiiiseet ee tee ete ae vee cn 1.062 1-375 313 28 (2) I- 1-68 a PN LOMA er acitercis nice ala'eisieis sac eins a e'v'eis eins atepiais aha eke a alo teis | a J es oy () 1.687 2-250 +562 33 16 1.750 1-937 187 10 (2) I. 2-500 +56: IND; 6; (sriale 4)... sij0s see aaapniqasssapss os acts acttpicacsbaea Teatanaecd ctl - he ae a6 2.500 3-125 -625 25 20 3-125 4-312 1.187 38 ox NOs ITI e ae a0 Ssi5.s «nic dap a eawe ea coco hls o-caebele wise eves nin vin ois | 4-250 5-500 I-250 29 (2) ONO RUE Ste foiet a iaia sicleimaiiea’s eres siclosretsers sin/ars asia eomaiars ater tie aiate(ars 5-500 7-000 I. 500 27 (2) 1.687 2.125 +437 25 (2) Noo, femases: ib. ssagsy = dadas tase sae rages oct eekes tp bciqaitays ees - 2-125 2-750 «625 29 7 2-750 3-750 1-000 36 ae Nuwixon female pple rw cara ee ee ee eee oe eee Hee 5) 3-062 +562 aa (2) 3-062 4-062 1.000 32 23 Itoi Gases bao saa ane sce adrmoceepoas ao soupHabcsORosbEMSenenons { pee = ERr =. a4 () 1.687 1-937 +250 14 27 2.000 2-250 +250 12 (2) Napa Gs cia oe aepnield s wstgicinipnininse'd conse sine Seieiningle soap in Sc oje No record. ¢ Sex unknown. From the data here given, combined with those furnished by Binford’s results (Table A), it is possible to work out a table showing the number of moltings, the suc- cessive sizes of the stages, the percentages of increase, and the intervals elapsing between moltings of the crab while passing from the last megalops stage to the usual adult size. Such a table follows. LIFE HISTORY OF THE BLUE CRAB. IOI TaBLe C. Intervals Stage. Width. Increase. between molts. Inches. Inches. Per cent. Days. SIE aE Nea ara tenia sn reir oinss sets nde le vonetec ere fafa param base as oa ics urease ei AI arpa puta rasarati cal nsucattcamnv wae ora tarbrac wiaco cats ESMRSECEA Hotere ctrcre cisto(cpieeice tt nrc eiaside sires ciewavaiciaas tists ciearcatince aiapae emaceh -128 0. 088 FEW a essa taeitis SSEASISIEPRMORELLD 01, CIRM aie alc eee E vie Total time, 208 days, or 6.9 months. It will be seen that, according to this table, 15 moltings occur. It is probable that the number is not absolutely fixed, but that it varies somewhat with the individual crab, usually being about 15. The size and rate of growth of individual crabs vary so that not all reach the width of 7 inches. No doubt, however, as many moltings occur, on the whole, in the smaller specimens as in the larger, the individual being smaller at the start and the actual increase in size at each molting being less. Various factors, such as temperature, food, etc., certainly affect the growth rate. Ninety-six adult female crabs were measured, and the average width was found to be 6.117 inches. No similar records are available for the male, but the average width is probably about 6.5 inches at least. Individual specimens of males are found which are much greater in width than this. Two exceptionally large specimens were measured, one of which proved to be 8 inches in width and 20 inches from tip to tip of the extended claws, and the other 8.5 inches in width and 1.25 pounds in weight. Specimens of females 7 inches in width are occasionally seen. The average increase in width at each molting is 32.71 per cent, or about one-third. This does not include the change from the last megalops to the first crab stage, at which time the increase in width is over 200 per cent, owing to a material change in the form of the animal. The increase from the first crab stage to the second is over 50 per cent. At the subsequent moltings the increase varies from 22 to 44 per cent, usually being about 33%. The time elapsing between molts is less in the early stages than in the later, averag- ing 6 days during the first 4 stages, 13 during the next 6, and 25 during the last 4. The average for all the stages is nearly 15 days. The results above set forth are confirmed by data obtained from some unpublished notes which were kindly placed at the disposal of the author by Prof. Hay. Hay 102 BULLETIN OF THE BUREAU OF FISHERIES. observed the molting of 22 immature crabs which were confined in floats on the water at the United States Fisheries biological laboratory at Beaufort, N. C., during the sum- mers of 1913, 1914, and 1915. ‘These crabs ranged from 1.0625 to 6 inches in breadth. His results are summarized in the following table: TABLE D. Days Individual crab. Width. Gain. between molts. Mm. Mm. Md sisie.cinleo sq eiecerelalpBlocpieve\oie ShMatue'Sinia,e stu) kath ule e'b/ae's ate wlweraawinia/Gunfa weld oiaivaher iwi e alerelajeieaie wath weet Ga7to 35 8 15 _SHRASONESCoSauE | OCbsost, —Csarcute:® cCostosnne. APROCne Coroner mr rat are rs Srp rr ek tor! gito 38 Ir Bisa Riaeinince duin's,c: cOhlya wane cWatath e.cin'e w sine Sigh Qin cin osbimarersiaa aad dhiine aisles oarhdeare aida) Aten bersts silage grto 38 7 ir AS cist orvia.e'aediu.s is ca Q@R we nunca apteiptore'e’s a(cie eral b niaihic s/s 4.625 to 6 inches. It will be seen from this table that the gain varies from 7 or 8 millimeters in the first stages to from 26 to 37 for the last. The gain is about one-third in 15 of the cases and one-fourth in the other 7. The interval between moltings varies from 11 days for the earlier to about 3ofor the later stages. The average for the first 11 cases is 16 days and for the last 11 cases 28 days. The total time required for the hypothetical crab of Table C to reach maturity from the megalops stage is 208 days, or nearly 7 months. Allowing a month (see Bin- ford’s results) for the completion of the zoéa and megalops stages, it will be seen that about 8 months would be required for the crab to reach adult size after hatching. All the evidence at hand, however, leads to the belief that the crab does not molt during the winter months. Perhaps the best evidence for this is the fact that, even before the close season was established in Maryland, the soft-crab industry ceased some time in October of each year until the middle or last of the following April. As this industry is dependent on the securing of molting crabs, they either do not molt during the winter or retire for the process to water of such a depth that access can not be had to them by the scrapes used in securing molting crabs nor by the oyster dredges. LIFE HISTORY OF THE BLUE CRAB. 103 Adult crabs which were kept in floats were found to become sluggish and to take no food after the temperature of the water fell to about 50° F. A juvenile crab, No. 6 of Table B, was secured on February 25 and kept until April 20 in the laboratory, where the temperature of the water in the aquarium varied from 44° to 55° F. No moltings occurred. Other experiments, carried out during the summer, proved that crabs will molt freely in aquaria in the laboratory. Binford’s work also shows the same to be true. On April 20 this crab was placed in a float in the water of the bay, the temperature of the water being 60° to 65° F. On May 5 thecrab molted. Very probably crabs do not molt during the season when the temperature of the water is less than 60° F. From temperature records of the water of Chesapeake Bay, kept throughout the year, it has been found that between about the last week of October and the middle of April the temperature of the water is below 60° F. The bulk of the young crabs of Chesapeake Bay hatch during the last two weeks of June and the earlier part of July. From that time until the last part of October four months elapse. During this time the crabs pass the first two stages and reach probably about the ninth or tenth crab stage, attaining a width of about 1.25 to 1.50 inches. Then come the winter months, during which time the crabs most probably do not molt, but lie dormant on the bottom. Growth and molting are resumed about the middle of April or the first of May. During the next three and one-half or four months the crabs molt five or six times and reach maturity during the last part of July or in the month of August. This agrees with the fact that during the six weeks from the middle of July to the last of August most of the pairs of mating crabs are found. As is described more fully on page 104, this occurs in the female at the time of the last molting and is thus a sure index of her arrival at the adult stage. The best evidence we have, then, points to the probability that the crab reaches the adult stage about 13 or 14 months after hatching. If, for example, a crab is hatched during June, it will reach the adult stage and mate during the latter part of July or the month of August of the following year. After leaving the megalops stage the abdomen of the male assumes the character- istic | shape, which is found throughout the remaining term of his life. The abdo- men is broad at the line of attachment with the body, but curves in shortly to anarrow , portion (Pl. LIV, fig. 35), which lies in a groove in the middle of the lower side of the body. Plate LIV, figure 34, represents the adult male crab when viewed from above. The abdomen of the female, after she leaves the megalops stage, is broad at the base of attachment and tapers to a point, each side forming almost a straight line (Pl. LIII, fig. 32). It lies in a depression on the lower side of the body and is held quite firmly in place by a pair of hooks which project from the body into cavities in the sides of the abdomen. ‘This form of abdomen is found in each stage of the female until adult size is reached. At the molting from which the crab emerges as an apparent adult the abdomen changes to almost a semicircular shape, except for a small point at the tip (Pl. LIII, fig. 33). It no longer lies in a depression of the body, and there are no hooks. It is held against the body by the effort of the muscles alone. The swimmerets seen on the under side of the abdomen, when it is pulled away from the body, are large and conspicuous (Pl. XLIX, fig. 16). In the pointed form these are small and insignificant in appearance. Plate LIII, figure 31, gives a view of the adult female when seen from above. It will be noted that the body is relatively longer 104 BULLETIN OF THE BUREAU OF FISHERIES. from beak to abdomen than that of the male and that the claws are smaller. The adult male can usually be distinguished from the female by these characteristics with- out the necessity of examining the abdomen. It is most probable that this molting, in which the change in the abdomen is involved, is the last one which the female undergoes. Adult females were kept under observation for several months in crates and floats. None was observed to molt during this time. Immature crabs molted when kept under similar circumstances. In the region of Crisfield, Md., the center of the soft-crab industry, hundreds of thousands of crabs are caught shortly before molting and kept until it occurs, in order to secure the soft crabs for the market. The author examined 2,624 cast shells obtained from various ones of the floats in which such crabs are confined. Not one of these shells bore the broad abdomen of the adult female. No crabber was found who could recall ever having seen a cast shell bearing the broad abdomen. Females with broad abdo- men are virtually never found exhibiting the-easily recognized marks that distinguish a crab which is preparing to molt. Three crabbers were found who said they had seen one or two adult females bearing such marks, but that they did not molt when kept in the floats. It is very doubtful if the marks observed on the crabs in question were really the same as those that characterize a crab in the premolting stage. As no such change occurs in the abdomen of the male, we have no criterion other than that of size, general appearance, and the manifestation of sexual activity by which to judge its probable state of maturity or whether it molts again after reaching maturity. The average maximum size is, as stated above, about 614 to 7 inches. It is probable that the males become sexually active somewhat before attaining the maxi- mum size, although the evidence on this point is rather meager. Both male and female crabs are found, especially during the winter, whose shells are discolored and bear barnacles, oysters, etc., apparently giving evidence that the shell has not been cast for a considerable period of time. There is no especial reason to suppose that the male molts indefinitely, in contradistinction to the female, which most probably does not. MOLTING. As already stated, the crab increases in size only when it molts, or sheds its shell. It is not exactly true to state that the crab grows only when it molts or that it grows by molting. It molts because it has grown and the shell, being inelastic, is too small and is thrown off.¢ Thus there is a sudden abrupt increase in size due to the expansion of the organism which has previously been crowded and somewhat wrinkled up within the old, hard shell. The actual molting process has been described by Hay.’ His excellent photo- graphs illustrating some of the stages of the process are included in the plates appended to this paper. The account given in this paragraph is modeled to some extent after his description, although an effort has been made to present the matter in more detail than he employed. As the crab approaches the molting period it begins to show its condition by various external markings, or signs. The first indication is the appear- ance of a narrow, black line just within the thin outer and back margins of the two @ Herrick, op. cit., p. 200, b Hay, op. cit., p. 411. LIFE HISTORY OF THE BLUE CRAB. 105 outer segments of the swimming legs. In a day or two this line becomes white, and the crab is known as a green, fat, or snot crab. Within three or four days the line becomes pink or red. (See Frontispiece, upper small figure at left.) It is formed by the edge of the segment of the leg, which has become loose from the old shell about it and has produced a bright fresh one, which is visible through the thin, outer old shell. A crab bearing the pink or red line, sign, or ring will molt within two or three days and is known as a peeler or, more rarely, a sluffer. A set of fine wrinkles also makes its appearance on the blue skin between the wrist and upper arm of the claw. A reddish color begins to appear at the margins of the segments of the abdomen. The carapace, or black shell, and the top shell of the abdomen are naturally not continuous, and, when the moment of molting is at hand, the carapace begins to be lifted up slightly, so that a gap appears between it and the abdomen. Then, on the under surface of the carapace, a crack appears in the shell at each end of the gap just men- tioned. As the carapace rises the crack on each side lengthens, passes below the spines, and extends nearly to the mouth. The posterior part of the body begins to protrude through the gap thus made. A crab in such a condition is known as a “buster” (Pl. LU, fig. 25). At this time it usually lies motionless, but can swim quite actively if disturbed. The remainder of the molting process requires only a few minutes, usually about 15. The carapace is lifted higher, the swimming legs begin to be withdrawn by rhythmic throbbing movements, and the body protrudes more and more from the shell (Pl. LII, fig. 26). The claws are the last to be with- drawn from the old shell, and, as they are large, some difficulty is involved in pulling them through the narrow arms of the claws. To obviate this, a roughly triangular portion of the shell of the large segment of the arm breaks along the outer side and rises up like a flap (Frontispiece, figure at lower left, and Pl. LII, figs. 26 and 27). The opening thus made extends along the arm to the body, and thus the natural opening for the connection of the muscles of the arm with the body is enlarged. Through this the large claw is drawn (Pl. LII, fig. 28). Immediately after molting is completed the skin is soft and wrinkled, and the spines are curved forward (Pl. LII, fig. 27). Although the crab is flabby and apparently helpless, it is capable of walking or swim- ming slowly if disturbed. The spines soon become extended, and the shell fills out and begins to harden. A female crab was taken while in the buster condition and was placed in an aquarium in the laboratory, where it was allowed to complete the molting process. This crab measured 3.5 inches in width before molting. At 10.15 a. m., the instant after molting was completed, and at later intervals, the crab was measured. ‘he results are sum- marized as follows: Date. Condition. Hour. Width. | Inches. pte ace ye, th series «cpa Site as Stare vino sbpihrs w Medea’ « wierd es «Sia tp rads inte f A CLORERMOIEING, . Sexy io] +210 Aa Ee week 3-500 petted atin Gann aaa aoe edn as elation? apa snts’ ksi tin doebesaf nie Giarcls cleipie enue n= | SILSCRELOLCEGLS, oe acts caies 10.15 &.M..... 4.000 a taetecctateees aidlersiviata’sidieni Shell leathery, ‘“‘buckram.” ¢ Shell nearly hard. 106 BULLETIN OF THE BUREAU OF FISHERIES. It will be seen that nearly all the increase in width, which amounted to 1 inch, was accomplished almost within the first hour after molting, and that, within 414 hours, the entire increase was completed. Within 24 hours the shell was too leathery to admit of the crab being used as a soft crab. A crab with such a leathery shell is usually called a buckram. Within 48 hours the new shell had almost reached the usual state of hardness. Effect of the moon and tides.—It is a popular superstition that both the moon and the tides have a marked effect upon the molting of the crabs. It is supposed that at cer- tain stages of the moon more soft crabs, or peelers, or whatnot, may be found than at certain other times. In the course of the experiments which were carried on in connection with the molting of crabs moltings occurred on the following dates in 1917: May 5; June 1, 12, 13, 19, 24, 26, 29, 30; July 2, 4, 9, 11, 12, 17, 21, 27, 31; and August 13, 20, and 23. Inno case were there more than two moltings on any of the above dates, and usually only one. It would be difficult to establish any relation between the changes of the moon and the moltings observed in these cases. No evidence whatever exists for the belief that the moon has any effect upon the molting of the crab, and the matter may be dismissed as of a class with all folklore superstitions concerned with the supposed rela- tion between the moon and mundane affairs, such as the weather, gardening, and the like. In many places it is thought that there is a close relation between the rise and fall of the tides and the molting periods of crabs. Many persons claim that the young crabs molt at every tide or every two tides. This idea is shown at once to be erroneous, since at least 48 hours are required for the crab to reach the usual state of hardness after molt- ing. The experiments with molting crabs, here described, were carried on where there was a tide of at least 4 feet. Some crabs were confined in floats which rose and fell with the water, others in wire cages resting on the bottom. No relation whatever was found to exist between the movements of the tide and the molting of the crabs. The movements of the tides do affect the distribution of the crabs. Immature crabs especially tend to come in with the tide. Busters which thus come in with the water and soft crabs that have molted after coming in are rather inactive and slowly follow out the ebbing tide. For this reason the best time to find such crabs is on a “half tide,” as it is falling. This, and other similar facts, have given rise to the notion that the movements of the tides actually hasten or delay or in some way regulate the molting act. AUTOTOMY. Closely connected with the process of molting is that known as autotomy, or the automatic throwing off of the appendages of the body. This phenomenon is common among crustaceans and has been the subject of considerable research. Little work, however, has been done in this line in connection with the blue crab. If a crab is seized or held by a claw or leg, it often throws off the appendage and escapes. The break occurs across one of the segments near the body, there being an arrangement to prevent excessive bleeding. The crab is thus often enabled to escape with its life at the expense of an appendage. This latter loss is not always as serious as might appear as the power to regenerate the lost appendage is possessed by the crab, at least until the molting stages are completed. If the loss occurs shortly after a molting, the regeneration will be made at the next molting. If it occurs only a few days before a molting, the renewal takes place at the second subsequent molting. LIFE HISTORY OF THE BLUE CRAB. 107 In the process of regeneration a very small, white papilla, or protuberance, first forms in the end of the old stump. This papilla enlarges and becomes a sort of thin- walled sac in which the new appendage is formed. As the sac becomes larger the new appendage can be seen folded up within it. At molting the sac is thrown off with the old shell, and the new limb appears in its normal shape, but is smaller than the corre- sponding member of its pair. At subsequent moltings it increases in its relative pro- portions and eventually attains the normal size, unless the loss occurred only a molting or two preceding the acquisition of adult size. In that case the new appendage remains smaller than its fellow. Many adult crabs are found in which one claw is smaller than the other or both claws are below normal size. Such an undersized claw is termed by crabbers a ‘‘jew claw,” i. e., reduced or ‘‘jewed down.”’ As an example of autotomy, the record of a crab which was kept under observation will be presented. A female crab, No. 9 of Table B, 1.687 inches in width and with the left claw missing, was placed in a small cage in the water on June 1. No papilla or limb bud had begun to form. This crab was kept under observation until July 21. The growth at the moltings which occurred during this time is recorded in the following table: TABLE E. Lengthof claw,includ- | Width of claw in ing entire arm. widest place. Date of molting. Width. Right. Left. Right. Left. Inches. Inches. Inches. Inch. Inch. I tremerc seen ete facie creas cee riere aidwig'e atajsbiaeistc siete dlobiars » sis's'o(a[s 1.687 1.625 (») o. 18 (0) FUME eet ee ce oisie ces ae tanate alate eins «in mv’ ein e's wieyaiaipi e's pie\s vie a hippie mistalel 2-125 2.250 1-75 -25 o. 18 FROIN oy otee es orninw 5 nlo:n'o)s is nlnjnie n'e10'e(niais'0ininln sininipio niniololnia='alnin’n(>o.a afaia 2.750 2-930 1.87 +35 +31 SUR ea Ret teta fe nein ole vip ae wiainisisiety siptciainte alain s'a'w\nlala nidiniajeta‘e\siniala.v- aiaie\yidio.a 3-750 3-500 3-50 +50 +43 SEAM MNESS otne a ctoietada ac csleitaictaeebtdee Claw missing. ¢ Gain in 5 moltings or 51 days from time claw was lost. It was necessary to close the experiment at this point. The left claw had practically reached the size of the right in three moltings, or in about 51 days from the time when it was lost. Plate LI, figure 24, represents the successive molts of the claws during the course of the experiment. In the juvenile crab the completion of the process of regeneration never occurs except at the time of molting, although limb buds may be seen forming before this. The presence of a fairly large-sized limb bud is a sign that the crab is approaching or is already in the peeler state. If, as has been stated, molting does not occur after the crab has attained maturity and if regeneration does not occur except through molting, the adult crab can not renew cast-off appendages. Further research is necessary to clear up this point. Adult crabs cast off the appendages apparently as freely as do the juveniles. Whether they regenerate these or not, the author is not in a position to state. An adult female with limb buds has never been seen by the author or by 108 BULLETIN OF THE BUREAU OF FISHERIES. crabbers who were questioned. Various adult females, as well as large males, which were kept under observation lacked cettain appendages at the initiation of the experi- ments. During periods of from one to three months, throughout which the experiments were continued, the appendages were not renewed, no signs of limb buds formed, nor did any moltings occur. Quite large males, 6 or more inches in length, have sometimes been observed to bear limb buds. As males, however, are found which are at least 8 inches in width, it could not be said that the individuals with the newly forming appendages had yet reached the adult stage. Large males with discolored shells, barnacles, and similar apparent evidences of age, have not been observed to have appendages in the process of formation. It can not yet be stated definitely whether the adult crab has the power of regenerating appendages removed voluntarily or involuntarily. MIGRATION IN CHESAPEAKE BAY. The migrations of the crabs found in Chesapeake Bay are of sufficient interest to merit special discussion. Nearly all the sponge-bearing crabs are found in the southern part of the Bay, in fact, far enough south so that very few occur in Maryland waters. The chief spawning grounds are in the waters of the lower part of the Bay. Records kept by a leading crabbing firm of Hampton, Va., for the summers of 1906- 1913, inclusive, show that the average number of male crabs was 11.8 per cent of the catch handled by this firm during those seasons. During the summer of 1917 the author kept records of the percentages of male and female and sponge-bearing crabs found in certain lots taken at random from the catches brought in to the crab dealers at Hampton, Va. The results are summarized in the following table: TABLE F. Soo Sa : earing : : earing Date. Melee crabs in Date. Malem crabs in 4 entire ca. | entire eatch.a catch.a Per cent.| Per cent. Per cent. | Per cent. PETE 2) fant oslo hin sagt EDAD SAR OCEe DESHE Cee a cree eer 142 PRRaaer ta Lye EAL EL A Cer retest) case state ces sols) chores hd nha batoa s, phaeta cease agar eseyereteicherjayakelasaie:eieye! agai evale, sre: osbie-cpsisisriye 142 ANSE OL Salil Be pet neOlisc Boece De OSnO See Faber oo Seco Inne Een OBB Er aan am 142 IPR ULAVLE TET ES pM TGD sco fe gaye este le ces cyosetciy eusueten Seeseiey® 0a ciclo acy ages tare se kaebale wipieve sts eves fit 142° RoI ORO MEA KONE | LC Aeer eee Meme seretete erases acer Meir ee Meee ce oe ote or meee see oe 145 Hamel yatta ploscleti casey verde. othe bic cetels velers, le aicwetnrs sacle che aoc Hoetiolala, ca atid Std Sharatinerets TRaR I 145 Sa pkanilyeRenierina me nig ananei igs ere ty ters! eerie toh gota Seen high le sis e's 145 EVeTI Sh ag NG OSE ys A Soe EE tans Uva scioeauccn carps cots ovietyamke eupse ete sam areamer 145 13s, TOES GATE, OSC Col Shaae A OO ROIS SOE EEC OBE Gee BS eet Serre AEE eIoRIE OIE ran tas 145 IATA yaU SE ACTOORN cement scscerits riers antics cle sextet ttre Maa nic Staqeinocgae eisidiecwiereleie sii vie eiece elsitg eS « 147 SyHio eh ond ip AG feet hee 5 al a bona oe eae i ola iene AU nth at Aedes Gahan Sani 147 capers eel bE A or Gera i be as LN kn eS eS 147 Sapelophs le sone... faerekeeers «oa Re, tks Lele MOREL ee a aed eho Logs «% 147 PB SPETIO PSS Sal Lela aewian ts ks kia Shiseido a= cr spretacainatenspans Ohta aie Sorag stays 148 SFR OUST OIE ISIN ote eladnte oi sia os hegae os RAT apy Nia ANE ye ayant sies spi Pais, isip ses, Sache 148 issn end tye l OPM tesa. orl erste a sues aleeiatctare edatere Sane ele it nialaie es igs e/eeayste araietelst tage 150 TE COTOUMENSTS NULISOMLNE Ec iat acierete ner: Sena ec ceen ea ee came sete ae eocaetae oes 150 Subfamily. Phlacodickyinss. Feet. hk. SE, BO EL neta. cas cans teee HS ONER. ah kos 152 Philccodietyo mi Garten ejay. exe. eso seer steny =p iethtaveds:- 0 sue ciel SANS cto teeter, 4 152 JER TOTS ma orS SE TB ete 4B DADE Clr abe Tab ACoA ee wee AOE Copue Cor oe 152 Sabtataily PHoriospor ei eteey aca sss cloysdnieseyayelaravale prea ebulsy aie f=, aa) ptniarsyeia: elevate oher'ai ais afevmcarerguerat 153 Inne mlos aoe CHE ELAS a cbtdn othe Baca dbo D bpon ye OF eEEMe Sone aar ono Tom banaar 154 NG SUM TETUSYS ILE SPR a Star ctere ee ci eesrttste facet eyerete iets aieie im aitovee eso oie srerere esi nye SCeaaLOre 154 Sipramilly: Bebyomitice wanker, CALE Mace aodee cea ea SER ha od satan en wien 157 IMicractonasbomgeniartice . Hictirnaeeiebtotess selec Ret obelee MSERES Skala setae = 6c 161 132 CONTENTS. ROP ALOSA, ©. «5 sycievaid wince ie male otetior a tetepe: sare aeneere arn aver senate hayepoy Mets ake enats (ovens atenatolia ate fore va) ove’ (ata S)s ualele Bamily, Darwianelldas(Aplysilitd asactey) | dererce cy eeteteveict cae aetna iniel ra feleiessile< sie/iays,icirie' eva cs Aplysillas WVE> Schitl7e (7:45 x c.oreseer teens ete PRT cis te lenoe acolecaints wins slorelecets oes AS LONGISP UE LISD fore ets welfare eho eh Aeie chevaR Re aah Peete lets cialoleisievee='« BamilycSpongelidea 223, oh cr, slats cteters onc yet nienesetote shasta eee eT tale ete crete Clea tate clea sto, Pleraplysilla Vopsent sc)... So e3. Sk eo ke Fe A ee ce ey elec eine eh ie owls PE Latensy TUS SP. arate cicesiey onesayore ei ayeteie Ta Stender etd Te eRe TLS eT TAS corer ahi ore 2 [Here yighs) a0) 0s CS PAHO CRMa HTC oun a nOMAgE “hn ama psukne Watuhoceousganhun aaasbeunane Supfamily, Stelospongiaae’. a2). - geienerose revelry oe « «lal ke se tele eee rari ele ee eetensuee hele tesa Hircinia Nardo. .s<.40026thtcnceesnae eed shea cnn enn he GR Sere OME eS RR eee Bibliographyc. cs: ect teas eek niece tee een pet hhn wh oh Seka RAW TNOe h Cem mem ema er SPeats Explanation) of plates: 2. <5 .chtscet tts ss cose teu GR anes cere Metre Imm Pant Meta teRte cle -1-j7- SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. & By W. C. GEORGE, Late Instructor in Zoology, University of North Carolina, and H. V. WILSON, Professor of Zoology, University of North Carolina. a Contribution from the United States Fisheries Biological Station, Beaufort, N. C. & INTRODUCTION. The following report includes a description of the forms of sponges which are pres- ent and in any degree conspicuous in the Beaufort area, The collections were made from time to time, especially between the years 1904 and 1907. Doubtless additional forms will be recorded. It is especially probable that intensive examination of ‘‘oyster rocks”’ and scattered shells will result in the discovery of other small, inconspicuous species, comparable in this respect to Pleraplysilla latens, herein described. Consider- ing the interest and value for experimentation of Pleraplysilla, a horny sponge of exceed- ingly simple character, such a search would be well worth undertaking. Collecting along the sea beaches has been incidental. Most of the material so collected proved unfit for precise study. Such sponges, in fact, are usually macerated, the microscleres lost, and the cellular tissues destroyed. The “Fishing Bank” off Beaufort Inlet has yielded specimens of four species. Collecting here, however, has been casual only, and the bank is probably the home of many more species. This bank is of a coralline nature, with a fauna which appears to be West Indian. It lies about 20 miles south- west of Beaufort and has been charted by the U. S. Bureau of Fisheries steamer Fish Hawk. Of the 17 forms described in this paper, those especially available for biological investigations of an experimental nature are species of Cliona, Suberites, Tetilla, Ren- iera, Stylotella, Lissodendoryx, Microciona, and Pleraplysilla, representing chiefly the two great monaxonid groups, but including also a tetractinellid species and a horny sponge. It is quite possible that some of the other forms, especially the species of Spir- astrella, Esperiopsis, and Hircinia, might be made use of for such investigations. These species occur in some abundance on the ‘‘Fishing Bank" and perhaps nearer the inlet. With care, living specimens, or, at any rate, living pieces which would answer the pur- 133 134 BULLETIN OF THE BUREAU OF FISHERIES. pose, might be brought to the laboratory, or certain breeding experiments might be begun on shipboard, and the cultures handled in the laboratory later. The scheme of classification followed is, in general, that used by Dendy in 1905. Some alterations which seem to be advisable have been made. In the scheme as here adopted the larger groups of the noncalcareous sponges are as follows: Order 1. Myxosponcipa.—Simple forms without askeleton. Absence of the skeleton primitive (Halis- arca, Bajalus, Hexadella, Oscarella). Order 2. HeEXACTINELLIDA (Triaxonida).—With triaxonid, characteristically hexactinellid, siliceous spicules. Order 3. TeTRAxonIDA.—The characteristic form of spicule is a siliceous four-rayed sclerite, each ray representing a particular axis (tetraxonid or tetractinellid spicule). But in some groups these spicules have been lost. Suborder 1. HomoscLEROPHORA (Dendy, 1905).—Megascleres and microscleres are not yet sharply differentiated from one another (Plakinide, Corticide, Thrombidz). Suborder 2. ASTROTETRAXONIDA (Hentschel, 1909).—Tetraxonid sponges without desmas, char- acterized fundamentally by the astrose microscleres, which, however, have been lost in the evolution of some groups. Tribe 1. ASTROPHORA (Sollas, 1888).—With tetraxonid megascleres and astrose microscleres. Tribe 2. ASTROMONAXONELLIDA (Dendy, 1905).—Sponge body generally compact and mas- sive, sometimes approaching a definite shape, but also incrusting. Megascleres all mon- axonid, often radially, or somewhat radially, arranged. Skeleton rarely fibrous, not dis- tinctly reticulate, and usually without spongin. Microscleres, if present, are asters of *some form or other. Presumably derived from the Astrophora through loss of tetraxonid megascleres. Equivalent to Hadromerina, Topsent. Suborder 3. SIGMATOTETRAXONIDA (Hentschel, t1grr).—Tetraxonid sponges without desmas, characterized fundamentally by microscleres which are either sigmata or forms derivable from the sigma. But the microscleres have been lost in some groups. ‘Tribe 1. SIGMATOPHORA (Sollas, 1888)—With tetraxonid megascleres. Microscleres when present are sigmata. Tribe 2. SIGMATOMONAXONELLIDA (Dendy, 1905).—Megascleres all monaxonid. Skeleton very commonly reticulate or fibrous, with a good deal of spongin. Microscleres, when present, either sigmata or derived forms such as chelz. ‘True astrose microscleres are absent, except, possibly, in an aberrant species or two. Presumably derived from the Sigmatophora through loss of the tetraxonid megascleres. Equivalent to Halichrondrina auct. Suborder 4. Lirarstipa.—Tetraxonida with desmas. Order 4. Krratosa (Euceratosa, Dendy, 1905).—Skeleton made up of horny fibers. Without proper spicules. Absence of spicules primitive and not due to evolution by loss. Sand grains and other foreign mineral particles often aid in forming the skeleton, and in exceptional cases constitute its chief part. The families, subfamilies, and genera represented are defined in the text. In constructing these and the above definitions we have freely used the memoirs of Dendy, Lendenfeld, Lundbeck, Sollas, Topsent, and Vosmaer. A consideration of some com- parative data, falling for the most part under the head of variation, together with a discussion of the facts on which the genera are made and classified, has considerably lengthened the sections assigned to several of the species. This matter follows, in each section, the description of the species, from which it is more or less conspicuously set off. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 135 ASTROMONAXONELLIDA Dendy. Family SPIRASTRELLIDZA. Megascleres usually styles or tylostyles, sometimes diactinal. Asters of various forms occur, often forming an ectosomal crust. Spirastrella O. Schmidt. Sponge incrusting, or cushion-shaped with processes, or massive; or vase-shaped with large cloaca, in which case the incurrent and excurrent surfaces may be differ- entiated. Megascleres styles or tylostyles, or a mixture of the two forms. Micro- scleres usually present, and abundant, in the form of spirasters, but these spicules may be exceedingly scarce, or even wanting. Spirastrella andrewsii, n. sp. (Pl. LVI, figs. 3, 6, 7a, b; Pl. LXVI, fig. 49a, 6, c, d.) A specimen was trawled August r, 1914, in 15 fathoms of water by the Fish Hawk on the “ Fishing Bank’’ off Beaufort Inlet, Fish Hawk station 8199. Since 1914 the Fish Hawk has taken in her summer dredgings several very similar specimens in the same locality. Several specimens of the species, now in the National Museum at Washington, were taken by the A/batross off the Carolina coast at a depth of about 30 fathoms during the summer of 1885. The species has also been taken in Jamaican waters by Prof. E. A. Andrews. The striking characteristic features of the sponge are its large size, habitus, and the differentiation of incurrent and excurrent surfaces. The sponge is cylindrical, with large cloacal cavity. Living specimens occur that are high and vaselike. The dried specimens, which, doubtless, have all col- lapsed more or less, are comparatively low and cushionlike. The external surface of the sponge ig incurrent, the cloacal surface excurrent. The following description is based on the specimen taken in 1914. The sponge is cushion-shaped, 60 centimeters across and 30 centimeters high; cloacal cavity 30 centimeters across at the mouth, 20 centimeters deep. The color is dark brown at the surface, lighter within. Consistency in an alcoholic specimen is like firm, dense cartilage, sponge becoming woody on drying. Whole sponge is greatly excavated by canals which contain many shrimp. The lateral surface is closely studded with small incurrent apertures (fig. 75), which measure 1 to 2 millimeters in diameter. The upper surface around the mouth of the cloaca shows irregular areas of similar apertures. They are all actual openings, not closed by pore membranes. Beneath the ectosome of the lateral and upper surfaces the sponge is cavernous, with large canals 6 to 8 millimeters wide, or even larger, extending more or less radially into the interior. Several incurrent appertures, perforating the ectosome, lead into each canal. On the walls of these great canals are abundant apertures leading into surrounding small canals. The external surface of the sponge between the incurrent apertures appears to the eye compara- tively aporous. It is, however, dotted with abundant small,round subdermal cavities, for the most part 200 to 300 » in diameter, but as small as 80 » in diameter. The thin, dermal membrane covering these cavities is perforated by pores 4o to 50 » in diameter, one to a few (about 3 to 4) pores leading into each cavity. The membrane roofing in a subdermal cavity contains spirasters, but is free from megascleres, excepting such as project into it from the surrounding tissue. The subdermal cavities are produced into small canals which pass inward, as may best be seen in a series of thick tangential sections. The anatomy suggests that the large incurrent canals serve to carry water directly into the deeper interior of the sponge, while the external region is fed by small canals, some of which arise as branches of the larger and others of which arise between the incurrent apertures, as just described. It is needless to say that observations on the living sponge in this connection are desirable. The cloacal wall is studded with oscula, 4 to 5 millimeters in diameter and smaller, although the smaller sizes are obviously often due to partial closure (fig. 72). Oscula are numerous, the distance between fully open ones being less than the oscular diameter. Each osculum, which, with its surround- ing rim of “‘oscular membrane,’ forms a circular depressed area in the dried sponge, is the aperture of 136 BULLETIN OF THE BUREAU OF FISHERIES. a large efferent canal extending more or less radially into the sponge substance. The efferent canals in this specimen are somewhat smaller than, and not so close together as, the corresponding incurrent canals; but this is a matter of individual variation. Numerous very small subdermal cavities, covered in by dermal membrane, underlie the cloacal surface between the oscula. Some are circular and about 80 w in diameter; others are elongated and about 2zoopmacross. From them small canals, often about 80 u in diameter, pass radially into the interior. (Sponge being shrunken, such dimensions are, of course, far from what they must be when the sponge isexpanded.) ‘The areas of dermal membrane covering in the subdermal cavities are thin and in life are doubtless perforated by the (now closed) pores. In correlation, perhaps, with the completely closed condition of the pores and the contracted state of the sponge these areas now contain megascleres scattered tangentially. The inference to be drawn from the anatomy is that the cloacal surface is not an exclusively excurrent surface. Spicules.—Megascleres: (1) The characteristic spicule is a tylostyle (fig. 49c), smooth, slightly curved, about 350 by 8. The common range in length is 325 to 425 uw’ in thickness,7 to1z uw. The head may be evenly rounded, or it may be somewhat elongated and bear a constriction. Modifications of this spicule sometimes occur in the form of (2) tylostyles (fig. 49a), in which the apex is not pointed, but rounded, and (3) styles (fig. 49h). Microscleres: Spirasters (fig. 49d), 8 to 20 » long, of one or two complete ‘‘turns.’’ ‘The spines are short and conical, sharp in some spicules, truncated and blunt in others. The spicules oceur at the dermal and cloacal surfaces and in the walls of the canals. They are abundant, but not very abundant; nowhere do they form a continuous layer or “‘crust,’’ but every- where they are spaced well apart. The common range of length in the spicules at the dermal and cloacal surfaces is 8 to 12 uw. In the walls of the large canals they reach 20 yw in length. Skeletal framework.—The septa of spongs tissue between the canals are well filled with megascleres except in the region immediately around a canal. In places the megascleres lie crossing one another in all directions. But in many septa, both thick and thin, the spicules alt lie about in the same direc- tion and are compactly arranged, thus constituting a fairly distinct tract. Different tracts cross one another at various angles, connecting and branching, and thus give to the skeleton a fibrous appearance which is inconspicuous in alcoholic, but conspicuous in dried, material. In the canal walls the mega- scleres are strewn thickly and tangentially. At the dermal surface, while there are some tangentially placed spicules, the bulk of the mega- scleres occupy an obliquely radial or radial position, their points projecting; but they are not divided into distinct brushes. They form a continuous covering, which is only interrupted by the large, incurrent apertures and the areas of membrane over the small, subdermal cavities. At the cloacal surface also many megascleres lie radially and obliquely, projecting slightly. At this surface there is in this specimen a particularly dense skeletal layer, about 1 millimeter thick, in which the spicules lie in all directions. This is a detail which is not present in all the specimens. The first specimen of this species studied by us was collected a number of years ago in Jamaican waters by Prof. E. A. Andrews, of Johns Hopkins University. Some data concerning this specimen may be recorded here. ‘The characteristic size of the tylostyles is 420 4 by 144. The spirasters at the dermal surface are commonly about 14 » long and are more abundant than in the Beaufort specimen, forming a continuous crust. In the walls of the larger canals they reach a length of 20 yu. In the dried fragment of this specimen sent us for examination, the incurrent apertures of the dermal surface were for the most part filled with what were doubtless small anemones. Dendy records (1896, p. 252) that in S. papillosa R. and D., occurring in the neighborhood of Port Phillip Heads, Australia, “the surface is sometimes much infested by a parasitic actinozoan.”’ The habitus of the Jamaican specimen as recorded by Prof. Andrews is interesting: ‘‘Height of sponge, 2 feet; diameter, 11 inches; diameter of mouth, 4 inches; depth of cloacal cavity, 10 inches. Sponge stood upright on a reef 20 feet below the surface. In life the color was black or perhaps purplish black, very dark brown when dried. When alive, sponge was covered with peculiar small objects which seem to be actinians, partly embedded in the surface, each 1 millimeter in diameter. Sponge hard, smooth, compact.”’ The Albatross specimens in the National Museum “‘average 18 inches in diameter, 12 inches high; cylindrical in shape, but with deep cavity in top.’’ (Letter from Dr. Mary J. Rathbun of the U. S. National Museum.) A particular specimen measured 30 centimeters high by 40 centimeters in cross diameter; cloacal cavity 15 centimeters deep, 25 centimeters across at the mouth. The sponge (dried) SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 137 ishard and woody. Color, dark brown at surface; canalar walls not so dark; parenchyma whitish gray. Large canals 6 to 7 millimeters in diameter, up to twice that size, extend in from both dermal and cloacal surfaces. A fragment of this specimen was examined microscopically. The tylostyles had a distinct but only slightly developed head; measured 414 to 468 w by 11 to 14 m; were in general slightly and evenly curved, but sometimes bent more or less abruptly. The spirasters were commonly ro to 12 pm long. In respect to the abundance of the spirasters at the dermal surface, the specimen (fragment) proved to be intermediate between the Beaufort and the Jamaican sponges. The vaselike habitus with differentiated incurrent and excurrent surfaces has not hitherto been described in Spirastrella. The nearest approach is made by the type specimens of Spirastrelia (Alcyonium) purpurea, collected by Peron and Lesueur in 1803 in Australian waters and first described by Lamarck in 1815. ‘Topsent, who in recent years has reexamined these specimens (1906a, p. 3), thinks that one, at any rate, possibly represents a marginal fragment of a vasiform sponge. The two faces are different. One, which Topsent suggests may be the outer surface, is imperforate and bears radial tuberosities, while the other bears orifices visible to the eye about 1 millimeter apart. The interior of the sponge is porous but not cavernous. Vosmaer (1911) has shown that a great number of forms described from many parts of the world intergrade in respect to any of the points he has considered. He hence combines them all as one species, which he designates S. purpurea (Lamarck). It may be doubted if the name is well chosen; certainly not if Topsent’s interpretation is correct and Lamarck’s fragments belong to a vasiform sponge with differentiated faces. This would be very different from the remainder of those combined by Vosmaer, and a new name would thus be necessary for Vosmaer’s species. Certain gross anatomical points of resemblance between S. andrewsii and Poterion are obvious. These are the vasiform shape and the differentiation of pore and oscular surfaces. But the pore areas and afferent canal system of P. atlantica are quite differ- ent structures from the incurrent apertures and great canals of S. andrewsit, and the oscula and oscular canals of the two forms are likewise very different. The resemblance is only the gross likeness which results from the sponges independently acquiring the same shape of body and the same type of distribution of the incurrent and excurrent orifices. : While S. andrewsiz is a striking species in the matter of size, one species of the genus already is recorded that exceeds it. This is the great Hymeniacidon pulvinatus of Bowerbank (1872, p. 126), which Vosmaer (loc. cit.) merges into Spirastrella purpurea (tropus pyramidalis). This sponge, occurring at Calebert Quay near Belize (British Honduras), is a massive, sessile form reaching 8 feet in height. The oscula and pores are scattered over the surface. The sponge is cavernous with large canals. Position of the genus.—The spiraster of Spirastrella has generally been regarded as a modified aster, and the genus accordingly put in the Astromonaxonellida (Hadro- merina of Topsent). But Vosmaer (1909) has concluded that the spicules are spiral monaxous with spines, since the latter contain no axial canals, as do the actines of a true aster. Dendy (1916, p. 96), perhaps reasoning from this fact, transfers the Spira- strellide, and along with them the Clionidz and Suberitidae, to the Sigmatotetraxonida. Awaiting Dendy’s detailed reasons for the change, the families are retained in this paper in their old position. 138 BULLETIN OF THE BUREAU OF FISHERIES. Family CLIONIDZ. Astromonaxonellida that bore into and excavate molluscan shells and other cal- careous bodies. Cliona Grant. The complete spiculation includes tylostyles, oxeas, and spirasters. Of these elements one or two fail to appear (undergo atrophy) in certain species. Cliona celata Grant. (PI. LVI, figs. 2, 4, 5; Pl. LXVI, fig. 50.) Spongia sulphurea, Desor, 1848, p. 67. Cliona sulphurea, Verrill and Smith, 1874, p. 450. Cliona sulphurea (Desor), Leidy, 1889. Cliona celata Grant, Lambe, 1896, p. 202. Cliona celata Grant, Topsent, 1900, p. 32. (Synonomy here given in full.) A common sponge in Beaufort Harbor, especially occurring in oyster and clam (Venus) shells. The specimen figured was taken just below low-water mark from the edge of a little island. The sponge consists of anastomosing trabecule (PI. LVI, fig. 5), which lie in the body of the shell, and numerous projecting tubular papillae bearing pores or oscula. The sponge trabeculae completely fill the excavations in the shell, and in an old specimen the excavations occupy nearly all the space between the thin shell walls. The papilla may be extended a few millimeters, or may be retracted intotheshell. They are of two kinds: (1) Pore papillz, which, when extended, have a tubular stalk with a mushroom-shaped cap covered by a dermal membrane riddled with pores 15 to 35 mw in diameter. No pores were found except over the expanded end of these papilla. The diameter of the tubular stalk of the pore papillz in a preserved specimen is 1 to 2 millimeters and the diameter of the expanded end 1.5 to 3 millimeters. They connect with the trabecule in the interior of the shell. (2) Very similar tubular papille, each bearing a single terminal osculum 1 to 1.5 millimeters in diameter, instead of a porous cap. In the living specimen, observed in shallow aquaria, the oscular papille are found to be conical at the tip and are easily distinguished from the pore papille, with which they are intermingled over the surface of the shell. They are few in number as compared with the pore papille. Coarsely granular or spheruliferous cells (cellules spheruleuses) are exceedingly abundant, as in the sponges examined by Topsent (1900). Spicules—Tylostyles, smooth, slender, slightly curved, with a pretty sharp point, measuring 200 to 400 n by 4 to gu. The spicules taper slightly toward the head end, as well as toward the point. The curvature is in the upper (head) half of the spicule. Skeletal framework.—The skeleton of the sponge trabeculae within the shell consists of irregularly scattered, moderately abundant tylostyles. The skeleton in the wall of the tubular pore papille con- sists of a dense confused network of tylostyles, some of which have their points extending slightly beyond the dermal membrane. At the distal end of the papilla, where the pore-bearing cap spreads out, the reticulum of the wall breaks up into a system of loose fibers and trabecule, which extend upward at various angles, spreading out terminally in brushes to support the dermal membrane which covers the cap. The skeleton of the wall of the oscular papilla likewise consists of a dense reticular mat of spicules. As we approach the tip of the papilla the mat becomes less dense, and the spicules point more toward the tip, becoming arranged frequently in more or less definite plumose tracts. There is no spongin. At the base of the pore and oscular papille there is a sharp contrast between the dense skeleton of the walls of these tubes and the loose skeleton of scattered spicules in the trabeculze within the shell. This sponge, as is well known, may grow out of the shell which it has excavated and eventually form a free mass of large size. But the massive phase, which is common on the New England coast, where it may reach a diameter of about 8 inches (Verrill and Smith, 1874, p. 127), has not been observed in Beaufort harbor. Farther south, on the west coast of Florida, the free phase (Raphyrus griffithsii Bow) is known to occur, both in the common massive form and in a branched tubular form (!) (Carter, 1884, p- 207). SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 139 Topsent (1889; 1900, pp. 34, 55) has concluded that there is no valid reason for maintaining the American form (Spongia sulphurea Desor) as a species distinct from C. celata. ‘This, in fact, seems to be the case, although a careful and detailed compari- son of specimens from the two sides of the Atlantic would in all probability show certain constant, if minute, differences. The Beaufort specimens examined do differ from the European individuals (comp. Topsent, 1900) in the following points: (1) The inhalent papilla are tubular except at the very top, where the trabecule which pass out from the wall to support the pore membrane encroach upon the axial cavity. In the European sponges (Topsent, loc. cit., pp. 35, 47) these papill are filled with tissue except basally, where they are hollow. (2) In the Beaufort specimens the head of the tylostyle only rarely exhibits a distinct apical prolongation, whereas this is the rule in the European sponges (Topsent). Topsent (loc. cit.) gives the range in size of the tylostyles as 180 to 360 4 by 3 tog. The range in size for the Beaufort specimens is close to this. Moreover, the shape of the spicules is the same, except for the above-mentioned detail, in the two sets of specimens. In the European sponges this is the only spicule that usually occurs. In very young specimens, chiefly in the papille, Topsent finds, however, spinose spirasters. But . these spicules soon cease to be formed. No such very young specimens have been studied on this side of the Atlantic. In some European individuals of this species long, slender, smooth oxeas, generally in fascicles, occur. Sollas has grouped these as var. linearis, but Topsent thinks the point is only a character such as separates indi- viduals and is not the mark of a subgroup. Poterion Schlegel. Beginning as a boring sponge, the body becomes free, large, and vase-shaped, with the incurrent apertures on the outer surface and the excurrent apertures on the inner or cloacal surface. Skeleton made up of tylostyles. Poterion atlantica,n.sp. (Pl. LVI, fig. 1; Pl. LXVI, fig. 5ra, }, c.) A single specimen was trawled by the Fish Hawk on the “Fishing Bank’’ off Beaufort Inlet at a depth of 14.5 fathoms. The sponge is vasiform, about 12 centimeters across at the top. The vasiform cavity extends entirely through the sponge, which, however, has had its base torn off. Actual height of the specimen is 11 centimeters. The height of the uninjured sponge was probably considerably greater. The outer surface exhibits contiguous, or nearly contiguous, circular, or irregularly rounded areas about 5 millimeters in diameter. These in the preserved specimen are slightly depressed. The central and greater part of each area is porous and recticular, as seen with the lens and even with the eye, this part measuring about 3 millimeters in diameter. These are the pore areas. Micropscopic preparations of the surface and sections through the pore areas show that each area includes numerous pores 75 to 100 in diameter. From each pore a canal of about the same diameter passes vertically into the cortex. The inner surface exhibits similar areas, the center of each occupied by an osculum o.5 to 1.0 millimeter in diameter. The osculum is the aperture of an oscular (chonal) canal which passes vertically through the cortex. The sponge has a gray, dense, cartilaginous cortex both on the outer and inner surface of the cup. The surface is now (in the preserved specimen), blackish brown, interior yellow. The interior looks fibrous and is comparatively solid. Spicules (P1. LXVI, fig. 51a, 5, c)—Smooth, slightly curved tylostyles, 210 to 460uby 4to8u. The head of the spicule may be globular (fig. 515), or there may be a slight constriction around it (fig. 51a), or the enlargement may be located a slight distance from the end (fig. 51c). There are no microscleres. 140 BULLETIN OF THE BUREAU OF FISHERIES. Skeletal framework.—The skeleton of the choanosome (Pl. LVI, fig. 1) consists of irregularly scattered megascleres, together with loose spicule tracts. Collections of sand grains occur abundantly in the choano- some. ‘There isa cortical layer 1 to 1.5 millimeters in thickness, composed of compactly and confusedly arranged spicules, which however, in the main, point in a more or less radial direction, sometimes ex- tending beyond the surface. This cortex is pretty definitely marked off from the underlying choano- some, in which the spicules are not nearly so abundant. ‘There is no noticeable difference in the size of the spicules of the cortex and those of the choanosome. At the base of the cortex there is present in most places a thin layer of spicules arranged more or less parallel to the surface (Pl. LVI, fig. z). The canals are in general small, mostly 50 to 150 nin diameter, but some are 500 » or more in diameter. The sponge tissue is dense and granular. The flagellated chambers are inconspicuous and measure about 30 4 in diameter. This interesting sponge is evidently very close to the well-known ‘“‘Neptune’s Cup,” Poterion patera (Hardwicke), of the Pacific, which Vosmaer some years ago (1908) showed to belong in the Clionide. Topsent reviewing Vosmaer’s paper (1909) would delete Poterion, merging it into Cliona. But the final structure assumed is such a marked one that the genus should be retained, as Vosmaer more recently has held (1911, p. 3). Poterion patera, which is not uncommon in the Malay Archipelago, reaches a height of 1 meter, with an aperture of 30 centimeters, the wall of the cup 25 millimeters thick (Vosmaer, loc. cit.). The Beaufort sponge differs from the Pacific species in the larger size and closer grouping ofits poreareas. These (Vosmaer, loc. cit.) in the latter form are indistinct in outline, something over 1 millimeter in diameter, and about the same dis- tance apart. The internal skeleton is stronger in the Pacific species than in the Beaufort form, consisting in the former of a firm trabecular network, the trabeculee made up of closely packed tylostyles and including in the axial region here and there a little spongin (Vosmaer). The spicules in P. patera range in size from 450 uw by 14 to 11 w to 200 by 10 to 7m (Topsent, loc. cit.). Family SUBERITIDA. Megascleres tylostyles or styles. Microscleres absent or represented in some forms by centrotylote microstrongyles. Suberites Nardo. Body frequently massive, but it may branch or become covered with outgrowing lobes. Without mammiform papille. Megascleres nearly always tylostyles. No microscleres. The spicules diminish in size toward the surface. The superficial spicules project radially, and the skeleton, as a whole, may exhibit in some measure a radiate arrangement. Suberites undulatus, n.sp. (Pl. LVII, figs. 8, 9, 10, 11; Pl. LXVI, fig. 52.) Fairly common in the muddy pools left at low tide around “Green Rock,’’ in Newport River. Sponge (Pl. LVII, fig. 8), a speroidal mass, made up of a basal undivided portion and closely set lamel- late and narrow, ascending lobes into which the former is produced over its superior and lateral surfaces. A characteristic specimen (the type) measures 60 millimeters in height, with transverse diameters of 75 and go millimeters. It was attached below to the shells of live oysters. A few pieces of shell were incorporated in its basal portion, end some alge grew out from between the lower lobes. Color light gray. Sponge fairly firm; compressible and easily torn. The lamellate lobes are all more or less radial, but flattened in various planes. They are thus inclined to one another at all angles. Where they meet they are apt to fuse; this produces cavities closed below and around the sides, which extend into the sponge interior and open above between the SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 141 free portions of the lobes. The upper margins of the lobes are fairly sharp and have in the alcoholic specimen a dense, whitish appearance. These margins are commonly notched and undulating, but the portions between the notches may be produced into ascending lobules. The central lobes are the longest; their radial length is about one-half the vertical diameter of the whole sponge. Probably the basal part of the sponge mass has been built up during growth through the gradual incorporation of lobes. The surface when examined with a lens is seen to be minutely roughened and well covered with projecting spicules. The ectosome includes very numerous subdermal cavities, varying considerably in diameter from about 150 to 1 millimeter (Pl. LVII, figs. 10,11). The thin, dermal membrane roofing these over is perforated by pores 20 to 4o » in diameter, several opening into each cavity. Most of the pores are closed, but enough are open to show the arrangement. -Smali oscula about 1 millimeter in diameter occur on the upper margin of the lobes. Probably they are naturally numerous, but now for the most part closed. The interior of the sponge is dense as compared with the ectosome, but sections (Pl. LVII, fig. 10) show that it, too, is greatly excavated by canals, most of which are small, about 100 to 300 uw in diameter, with some largerones. Flagellated chambers, ellipsoidal and 20 to 25 «1 by 30 to 35 u, are abundant in the choanosome. The thin trabecule and sheets of sponge tissue are favorable for histological study. ~ Spicules.—Tylostyles smooth and slightly curved, with well-developed, rounded head (Pl. LXVI, fig. 52). The shaft is very slightly thicker in the middle than near the head end, tapering at the other end to a sharp point. The head is not infrequently irregular, sometimes constricted near its middle. Range in size for whole sponge, 200 by 6 » to 460 by rou. In the interior the larger sizes are abundant, perhaps predominate. The spicules of the dermal skeleton are, in the average, smaller; the common range being 200 by 6 u to 320 by 8 n. Skeletal framework.—The skeleton (Pl. LVII, figs. 9, 10, 11) is made up chiefly of abundant and fairly compact tracts of megascleres, which pursue a rather vaguely radial course in the basal part of the sponge, becoming distinctly longitudinal inthe lobes. In aslice of some size through the basal part it is easy to see that, while many individual tracts curve in all directions, the skeleton as a whole ex- hibits a radial arrangement. The spicules lie more or less longitudinally in the tracts and are abun- dantly scattered between the latter. Spongin is absent. From the internal skeleton short tracts are given off which extend outward, usually upward and outward, through the ectosome and terminate in dermal brushes of divergent spicules (fig. 11). The dermal skeleton includes, in addition to the brushes, a good many single, radial, and projecting megas- cleres and abundant tangential megascleres scattered without order. The spicules of the dermal brushes usually project a considerable distance, often about half the length of the spicule. The spots at which they project are, as a rule, either not elevated, or only slightly elevated, over the surface in general; but in places these spots are elevated high enough to be called ‘“‘conuli.’’ The difference may in part be due to contraction, In the upper margins of the lobes, or of the subdivisions of the same, the dermal brushes are so closely set as to form a continuous furze, in which the longitudinal skeletal tracts terminate. It is this dense aggregation of dermal spicules that gives to these margins their whitish appearance in the alcoholic specimen. The lobes of the Beaufort sponge are, of course, structures quite different from the papille of the Polymastide. In its extensive development of the ectosomal canal system S. wndulatus resembles the species grouped under Topsent’s genus Pseudosuberites: P. sudphureus (Bean), P. hyalinus (R. and D.), P. andrewsii Kirkp., P. exalbicans Tops. But this particular feature does not, it seems to us, constitute sufficient ground for excluding the sponge from the older genus. Probably when the canal system of the numerous Suberites species has been studied more extensively, considerable variation will be found in this matter within the genus. 110307°—21——10 142 BULLETIN OF THE BUREAU OF FISHERIES. SIGMATOPHORA Sollas. Family TETILLIDZ Solias. The characteristic megascleres are protrienes, which may be very slender, arranged radially. The skeleton in general is usually strongly radiate. Tetilla O. Schmidt. Typically the ectosome is not a distinct layer, but shades off into the choanosome; pores and oscula scattered and not located in special depressions. In some species, however, the ectosome is to some extent histologically differentiated and partially assumes the character of a fibrous cortex; and in some species there are special depres- sions on the floor of which the pores and oscula are located. There is no special cortical skeleton. Tetilla laminaris, n. sp. (Pl. LVIII, fig. 14; Pl. LIX, fig. 17; Pl. LXVI, fig. 54a to h.) Fairly abundant in Newport River in the vicinity of ‘‘Green Rock.’’ The specimens used in pre- paring this paper were dredged at half tide, at a depth of 4 feet. Sponge body (Pl. LVIII, fig. 14) a vertical lamella, elongated horizontally, the lower part of the lamella rooted in muddy sand by abundant fascicles. The lower edge of the lamella is thin; from this edge the body thickens gradually to the upper margin, which isrounded. The lamella is sometimes folded; the folds vertical. Sponge dense, firm. Color in the fresh state, grayish brown. The root fascicles arise from the whole lower edge and the neighboring parts of the lateral surfaces; the uppermost, relatively high up on the lateral surface, are short; they increase in length toward the lower edge. In the collected specimens the length of the lower rootlets is for the most part 10 to 20 millimeters, but in one specimen the length reaches 50 millimeters. The rootlets are so abundant that the whole lower edge of the collected sponge bears, even after washing, a continuous mass of sand held in place by the root spicules. The rootlets were in large part removed from the specimen photographed. In the type specimen the length is 115 millimeters, the greatest height 60 millimeters, greatest thickness 13 millimeters. Smaller and larger specimens are common. The largest specimen in the collection is 180 millimeters long, with a greatest height of 70 millimeters and greatest thickness of 30 millimeters. Relatively shorter and higher specimens occur, but the horizontal length is character- istically considerably greater than the height. The surface of the upper part of the sponge body looks smooth to the eye. In reality, as may be seen with the lens, slender megascleres everywhere project from it for a fraction of a millimeter. Numerous small oscula, 0.5 to 1.5 millimeters in diameter, the apertures of short oscular canals, are scattered along the upper margin at intervals, 2 to 15 millimeters apart. A few occur in some specimens on the lateral surfaces, near the upper margin. Pores about 30 to 60 » in diameter abundantly scattered between the projecting brushes of spicules. They perforate the very thin dermal membrane and lead into small subdermal cavities which occupy an ectosomal zone about 60 to 80 » thick. The intact surface appears dense to the eye; with a lens it is seen to be finely diversified by the minute subdermal cavities. . The ectosomal zone and the whole peripheral region to a thickness of about 0.5 millimeter is denser than the interior, owing to the smaller size of the canals; but, while the canals of the interior are numer- ous and larger than those of the ectosome, they are only a fraction of a millimeter in diameter (Pl. LIX, fig. 17). No part of the ectosome is differentiated to form a fibrous layer. Skeletal framework.—The mesial region of the sponge lamella includes a number of spiculo-fibers which pursue, in the main, a vertical direction. From these, radial spiculo-fibers extend outward, terminating in a layer of closely set, peripheral, radial brushes, about 800 to 1,000 y in radial length; the spicules of the brushes, projecting for the most part a short distance, about 100 »; some of the pro- trienes three times as far (Pl. LIX, fig. 17). In the lower half of the sponge the radial spiculo-fibers pass obliquely downward. The rootlets are the prolongations of some of the radial fibers and of some SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 143 of the lower mesial fibers. The spiculo-fibers are compact, cylindrical tracts in which the spicules are arranged longitudinally, without spongin. Between the fibers are scattered megascleres. Spicules (P1. LXVI, fig. 54a to h).—{1) Skeletal oxea, smooth, equi-ended, tapering gradually toward each end. ‘There are two types which intergrade. The shorter form (fig. 54a) helps to make up the spiculo-fibers and is scattered between them. It is also abundant in the root fascicles. It is often slightly curved. Common sizes are 600 1 by 10 4 to 1,000 7 by 12 4. A longer form (fig. 545), with very slender extremities, generally straight or nearly so, with the ends sometimes curved or bent, is abundant in the spiculo-fibers. Common sizes are 1,500 » by 16 pz to 2,300 uw by 20 pn. (2) Oxea of the peripheral radial brushes (fig. 54c), inequi-ended, the outer end much the thicker; about 800 to 1,000 yw long, 8 thick near outer end, thence tapering gradually to inner end. In each brush there are several of these spicules. (3) Protriznes of three types. Very slender protrienes, with hairlike cladi (fig. 549); the most abundant spicule in the peripheral radial brush; rhabdome near outer end generally about 1 » thick, 0.5 millimeter long; cladi much thinner, hairlike, 10 to 60 uw long; spicules projecting and covering whole surface of sponge, like fine hair. Immediately around an osculum these spicules are slightly larger than elsewhere, although the cladi are no thicker. Doubtless some of these spicules develop into the stouter forms of protriene, but they can not be regarded as young stages of a characteristic skeletal element, for they themselves are a marked feature of the skeleton. A stouter protrizne (fig. 54¢) occurs in some abundance in the radial bundles of the lower half of the sponge, projecting from the surface; it occurs also, but rarely, in the upper part of body. Monzne and diane modifications are present. Rhabdome 6 to 8 w thick near outer end, tapering gradually and becoming very slender, about 2 millimeters long. Cladi fairly strong, 30 to 48 uw long, about 4 u thick at base. Immediately around an osculum abundant protrienes of the type shown in figure 54f occur, the spicules projecting in the usual way. Rhabdome at the outer end is about 3 u» thick, thence tapering gradually. Cladi about 2 » thick at base, 14 to 40 uw long, commonly of unequal lengths, one cladus often considerably longer than the others. With these spicules are mingled the common, very slender forms (fig. 549). (4) Anatrizenes of the ordinary character (fig. 54d) are abundant in the radial fibers of the lower half of the sponge, the entire spicule lying within the body; rhabdome about 6 » thick near cladome, tapering gradually and becoming very slender, about 1,500 to 2,000 mu long; cladi about 32 yu long, strong, diverging less than 45 degreesfrom rhabdome. The root fascicles are largely made up of similar anatrienes, in which the rhabdome reaches a greater length, measuring in some isolated spicules as much as 3.5 millimeters. (5) Sigmata, giving the common C and S shaped appearances (fig. 54h), are abundant in the ectosome, including that of the root fascicles, and in the walls of the canals. They are about 12 w long. The surface of the spicule is slightly roughened, sometimes passing into a minutely spinose condition. Hyatt has described and Sollas redescribed (Sollas, 1888, p. 46) a Tetilla, T. gravata Hyatt, from our Atlantic coast (Buzzards Bay), which is, however, a distinct species from the Beaufort form, although Hyatt’s species perhaps extends southward as far as the North Carolina coast. At any rate, we have a number of specimens collected at Wrightsville, N. C., and Ocean View, Va., by R. Budd Chalmers, of Wilmington, N. C., which are certainly not far from 7. gravata, possibly representing a variety. Unfor- tunately, all of our specimens are beach specimens, and the surface has been rubbed so that the spicular details necessary for a precise comparison can not be made out. Discussion of the genus.—Recent writers are not in unison with regard to the use of this genus. Lendenfeld (1903) merges into it Chrotella Soll., and later (1906) merges Tetilla (+Chrotella) in Tethya (Craniella Soll.). Lendenfeld’s action is based on the occurrence of intermediate forms, which make it impossible to divide this group of species clearly into the genera recognized in Sollas’s scheme, which are based chiefly on the anatomico- histological features of the cortex. The intermediate forms unquestionably exist, but 144 BULLETIN OF THE BUREAU OF FISHERIES. Lendenfeld’s treatment tends to obscure the nice distinctions to which Sollas’s classifica- tion gives expression, and which should certainly not be lost sight of. Topsent in 1904 continues to use the three genera, Chrotella, Tetilla, Craniella. Dendy in 1905 (p. 89) uses but redefines Tetilla so as to include forms in which the ectosome is in part fibrous. His definition runs: “Cortex absent or feebly developed; no special cortical skeleton.’ This is one way out of the difficulty presented by the occurrence of intermediate forms, in that Tetilla is here made to include species that shade off toward Tethya (Craniella), and which certainly are intermediate. To be sure, another classifier using the same genera might include such or slightly different inter- mediate forms under Tethya, extending Tethya downward, so to speak, rather than Tetilla upward. Dendy also uses Craniella in 1905, and again uses Tetilla in 1916, in the sense in which he employed it in 1905. Row in 1911 uses Tetilla and Chrotella. Hentschel in r911 uses Tetilla, but in 1912 follows Lendenfeld and merges Tetilla (+ Chro- tella) in Tethya (Craniella auct.) As exploration goes on the number of sponge genera known to run into one another increases. Everywhere intermediate forms are found. We meet, then, very frequently the practical difficulty of finding the record of a known species or of deciding where to record a new species. If the genera exhibited a linear arrangement, we might have sharply defined genera alternating with less homogenous intermediate ones. But it frequently happens that the species of a sponge family fall into groups which shade off in all directions toward one another. In such a case, and it looks as if discovery would show that this is all but universal in sponges, the questions arise: Shall we give up any formal grouping of the species (it is of course not a mere question of names, genera or subgenera) ? Or shall we define all the species groups (genera or subgenera) in a compre- hensive, and therefore rather loose, way, which results in overlapping? Or shall we meet the difficulty by accepting some sharply defined and other lcosely defined genera? It is the latter method which is commonly employed, although not always explicitly, and no better treatment has as yet been found. Tetilla, the simplest, and therefore presumably the ancestral genus of the family, has been gradually enlarged in the practice of recent writers (Dendy, Topsent, Hent- schel) by the incorporation in it of species that depart in one direction or another from the central group of typical forms to which Sollas’s definition is applicable. Topsent, for instance, includes (1904, p. 97) T. longipilis, in which there is the beginning of a cortex, viz, an ectosome which is in part fibrous; the species having differentiated in the direction of Tethya. Dendy (1905, p. 89) includes T. hirsuta, in which there are a more or less fibrous cortex and surface depressions, the smooth floor of which is per- forated by pores or by oscula; a species with Tethya and Cinachyra-like features and which Lendenfeld (1903), in fact, lists asa Cinachyra. In the same paper (p. 91) Dendy includes T. anomala, in which the ectosome is pretty sharply differentiated from the choanosome, is fairly thick and to some extent fibrous, and ‘‘almost amounts to a cortex;” evidently a species approaching Tethya. Another species approaching Tethya has been more recently recorded by Dendy (1916, p. 105). This is T. barodensis, in which there is a well-developed dense cortex which is “‘perhaps to some extent fibrous.” T. cinachyroides Hentsch. (1911, p. 283) and T. limicola Dendy (1905, p. 93) also deserve mention as intermediate forms; in the anatomy of the peripheral canal system, Cina- chyra-like, although they lack the cortex of Cinachyra. Tetilla in this paper is accepted in the extended sense. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 145 SIGMATOMONAXONELLIDA Dendy. Family HAPLOSCLERID Topsent. Microscleres often absent; when present never chele. The megascleres are usually diactinal. Where the skeleton is made up of distinct spiculo-fibers, these are typically not plumose. Subfamily RENIERINZ. Megascleres oxeas or strongyles varying occasionally to styles. Skeleton reticu- late, or the spicules may be scattered without definite arrangement. Spongin absent or present in small amount; only exceptionally does it envelop the spicules. No micro- scleres. Reniera Nardo. The skeleton is typically a close, uniform, reticulum, each side of the polygonal mesh formed by a single spicule. Spongin usually at the nodes of the reticulum. The side of the mesh may, however, be multispicular, and long miultispicular fibers may develop. Reniera tubifera,n.sp. (Pl. LVII, fig. 12; Pl. LVIII, fig. 15; Pl. LIX, fig. 16; Pl. LXVI, fig. 5sa, b,c.) Reniera sp., Wilson, 1910. A fairly common species in the harbor. The best collecting locality is Newport River, close to the town. The sponge is scattered over the bottom and may conveniently be taken at low tide. The body of the sponge (Pl. LVIII, fig. 15) is of irregular shape and consists of a reticular system of anastomosing cylindrical branches varying in diameter from 3 to 8 millimeters. It is not soft, but quite fragile. The specimen figured measures 130 millimeters in length, 30 millimeters in height. Rising vertically from the anastomosing branches are numerous tubes, 2 to 10 millimeters high and 1 to 3 millimeters in diameter, bearing oscula at their tips. In some cases these oscular tubes fuse with one another where they come in contact. The walls of the oscular tubes are colorless, thin, and trans- parent; the oscula at the tips measure o.5 to 2 millimeters in diameter. The dermal membrane of the sponge is delicate and is perforated by numerous irregularly dis- tributed pores measuring about 50 u in diameter. The pores open into small subdermal spaces, which ramify in the meshes of the ectosomal skeleton and lead into a system of very abundant canals in the sponge body (Pl. LIX, fig. 16). The flagellated chambers are conspicuous in stained sections and are very numerous. They measure about 25 wu in diameter. The mesenchyme is granular and rather scanty. Color of sponge pink or reddish purple, varying to brown; color fading quickly in alcohol. Spicules (P1. LXVI, fig. 55a, 6,c).—Smooth, slightly curved oxeas measuring 125 to 170 u by3to 8uyu, the smaller sizes, doubtless, being young stages. The usual variants occur in the shape of styles (fig. 54b) and strongyles (fig. 54c). Skeletal framework (P1. LVILI, fig. 12, Pl. LIX, fig. 16) —The main skeleton (fig. 16) consists of a com- bination of fibers, reticulum, and scattered spicules. The fibers course longitudinally through the component branches of the sponge and are conspicuous. They are 30 to 100 w in diameter, 3 to 8 spicules abreast, the spicules parallel to one another. In the parenchyma between the spiculo-fibers are many scattered spicules. These are commonly cemented together with spongin where they meet or cross, thus giving rise to a vague and irregular, predominantly unispicular reticulum, There are also many free spicules. The dermal and ectosomal skeleton (figs. 12, 16) is a distinct unispicular reticulum. The meshes are commonly three sided but may be four or five sided. The Beaufort species departs from the typical Renieras, in which the skeleton is a unispicular reticuium, and falls in the large group of species in which special multispic- ular tracts are developed in the midst of a skeleton that preserves more or less the orig- 146 BULLETIN OF THE BUREAU OF FISHERIES. inal character of a unispicular reticulum. (Vide Topsent, 1894), p. 4; Dendy, 1894, p- 236.) The following citations may help to put the Beatifort form in its proper place in the immense collection of Reniera species. In R. simulans (Johnston) Schmidt there are multispicular primary skeletal lines (Topsent, r901a, p. 356; Bowerbank, 1866, p. 308). The same is true in R. dancoi Tops. (Topsent, 19015, p. 12). Among other species falling in this group may be men- tioned R. pigmentijera Dendy (1905, p. 143), R. massalis Carter, and several other species recorded in Dendy’s Catalogue (1894, pp. 236-238). Where the habitus of the sponge is tubular the multispicular tracts may form Jongi- tudinal fibers curving outward toward the surface, connected by secondary tracts 1 or 2 spicules thick. This is the case in R. scotti Kirkpatrick (1908, p. 62), and is more or less true of R. spinosella Thiele, R. implexa Schmidt (Ridley and Dendy, 1887, p. 15; Topsent, 1904, p. 244), R. utriculus Tops. (1904, p. 246), R. wrceolus Rathke and Vahl (Topsent, 1904, p. 246; Lundbeck, 1902, p. 35). In several species the habitus is that of an erect lamella. In these forms also the mul- tispicular tracts are longitudinally placed, and may be strongly developed, more especi- ally in the basal part of the sponge. This is true of R. parenchyma Lundbeck (1902, p. 37), R. foliwm Lundbeck (1902, p. 39), R. ventilabrum Fristedt (Lundbeck, 1902, p. 40). In some forms of more or less massive habitus the multispicular tracts have no regularity of arrangement, e. g., R. zoologica Dendy (1905, p. 143). Forms in which the skeleton is made up in part of a reticulum and in part of dis- tinct polyspicular fibers might be referred, following Topsent (1904, p. 243), to Clado- croce Tops. The common practice (vide Lundbeck, 1902, p. 51) of not separating these species from the other less modified ones is followed, however, in this report. In several forms the originally uniform skeletal reticulum is only retained at or near the surface, becoming looser and less distinctly developed in the interior. This is the case with the Beaufort species; with R. (Isodictya) crassa Bow., in which primary multispicular skeletal lines are developed (Bowerbank, 1882, p. 126); with one of the forms (Reniera species 8) recorded by Hentschel (1912, p. 411), in which this halichon- drine tendency is not counterbalanced by the differentiation of distinct multispicular lines. On the other hand, there are forms in which the original simple reticulum is sup- planted in the ectosomal region by a reticulum composed of multispicular fibers. The original reticulum may persist at the very surface as in R. semifibrosa Dendy (1916, p. 112), or may here break up in halichondrine fashion into scattered spicules, as in R. fibroreticulata Dendy (1916 p. 111). In both of these species there are also internal multispicular fibers, and, as Dendy points out, a transition is made to Pachychalina. Finally forms may be mentioned in which the skeletal reticulum departs from its primitive character in that all sides of all meshes become multispicular (Topsent, 18945, p- 4). SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 147 Family DESMACIDONIDA. The characteristic microscleres are cheloids (chele and modifications), but forms are included in which these spicules presumably have been lost during the course of evolution. Subfamily MYCALINA. Skeletal fibers, or spicular tracts, without echinating spicules and not markedly areniferous. The body has no fistular outgrowths. Stylotella Lendenfeld. Sponges of soft texture. Megascleres, styles in fibers or tracts and scattered. No microscleres. Stylotella heliophila Wilson. (Pl. LVIII, fig. 13; Pl. LIX, figs. 18, 19; Pl. LXVI, fig. 53a, b, c.) Stylotella heliophila Wilson, 1911, p. 13. The most abundant sponge in Beaufort Harbor; common on the bottom in shallow water attached to shells, also under wharves attached to piles, stones, ete. Habitus varies. Sponge incrusts the shell or other substratum and grows up in the shape of lobes. These may be quite independent of one another. More commonly the ascending lobes fuse where they touch, and thus a more compact mass is produced, reaching, but rarely exceeding, 100 millimeters in diameter. The surface is roughened by minute conu- lose elevations 14 to 1 millimeter high. Color orange, sometimes with a greenish cast. A typical speci- men is shown in Plate LVIII, figure 13. The oscula are mostly located at the ends of the vertical lobes and at the ends of tapering, more or less conical, outgrowths from the lobes. The pores, which in an alcoholic specimen measured from -20 to 45 » in diameter, are irregularly scattered in great abundance over the dermal membrane. The dermal membrane (PI. LIX, fig. 18) is translucent. Beneath it may be seen a richly developed system of conspicuous subdermal canals 3 to 4 millimeters and less in diameter. Spicules (P1. LXVI, fig. 53a, b,c) —The only spicules present are smooth styles, slightly curved or sometimes straight. Therangeofsizeis120to350nby4togu. Inaddition, there are present some very slender styles measuring 115 to 225 u by 2 wor less. These are scattered in the parenchyma and are doubtless young stages of the skeletal spicule. * Skeletal framework (P1. LIX, figs. 18, 19).—The spicules of the interior are irregularly scattered. Here and there they cross one another so as to give rise to meshes, or they may combine to form spiculo- fibers or tracts (fig. 19). A small amount of spongin is present in the spiculo-fibers and at some of the points where the spicules cross. The spicular tracts are commonly present in the trabeculae between the larger canals. They often fray out in a brush-like fashion at the surface. In the ectosome are abundant, more or less radially arranged styles, some slightly projecting. In the dermal membrane the styles very generally project more or less radially, frequently form- ing the dense brush-like groups referred to above, but between these they are scattered more or less tangentially (fig. 18). Stylotella Lendenfeld was diagnosed by its author (1888, p. 185) as follows: ‘‘ Hete- rorrhaphide of very soft texture. Megasclera styli, in bundles and scattered. No microsclera.” Dendy (1896, p. 231) deletes the genus as not distinguishable from Hymeniacidon Bowerbank, which he places in the Axinellide. Topsent (1899, p. 109) retains the genus and thinks its relationship is with Esperella (Mycale). He gives the following diagnosis: ‘‘Esperelline with reticular skeleton. Fibers (at least the primary ones) multispicular. Megascleres: styles. No microscleres.’’ Lindgren (1898, p. 313) follows Dendy, merging the genus in Hymeniacidon. Kirkpatrick (1900, p. 137) retains the genus. Topsent (1904, p. 224) criticises Dendy’s treatment of the genus and retains it, placing it in the Esperelline. Dendy (1905, p. 185) again records his opinion that 148 BULLETIN OF THE BUREAU OF FISHERIES. the genus is not distinguishable from Hymeniacidon, which he continues to place in the Axinellide. Hentschel (1912, p. 355) retains the genus and follows Topsent in placing it in subfamliy Mycaline (Esperelline auct.) in the Desmacidonide (= Poeciloscleride Tops.). It seems to the writers that Topsent’s treatment is the correct one. Esperiopsis Carter. Habitus varies; incrusting, amorphous, and more or less upright forms occur; the latter may be leaflike or subcylindrical and branching. Spongin commonly pre- sent, the amount varying. Skeleton varies from a state in which there are well-developed spiculo-fibers, with abundant spongin, to a renierine or halichondrine condition. Mega- scleres, styles, some of which may undergo the strongylate modification, often with tylostyles. Microscleres, isochele, which may be accompanied by stigmata, toxe, or forcipes. Esperiopsis obliquan.sp. (Pl. LX, figs. 20 to 23; Pl. LXVI, fig. 58a, b, c, d, e, f.) Five specimens; three collected on Fort Macon beach; one dredged just outside Beaufort Inlet; one dredged on the “‘ Fishing Bank”’ off the inlet. Sponge is ramose; the branches cylindrical or subcylindrical, smooth or knotty, sometimes dis- tinctly compressed, commonly 4 to 6 millimeters in diameter. The main branches arise from a base and themselves branch. Fusion takes place sometimes between contiguous branches. Sponge may be vertical, or the branches may extend out in various directions from the base. The upright speci- mens range in height from 60 to 200 millimeters. A specimen with spreading branches has a greatest diameter of 50 millimeters. Sponge firm, but compressible and elastic. Color bright red. The known specimens show three fairly distinct types of habitus. But as no definable skeletal peculiarities are associated with these differences of the exterior, the types are, doubtless, only indi- vidual forms, reached as a result of particular growth and differentiation responses that are called out by the local environment. In one type (Pl. LX, fig. 21) the habitus is chaliniform. The branches, in general, are long, slender, cylindrical, smooth, and taper terminally. In one of the two specimens of this type a few lobes are slightly knotty here and there. Inasecond type (Pl. LX, fig. 20) the distin- guishing features are the knotty charactemof the branches and the biseriate arrangement of the oscula. In a third type the branches are smooth but enlarged terminally, clavate or spatulate. This type is represented by the specimen with spreading branches and less well by a small specimen 60 millimeters high, which, perhaps, was vertical. Dermal membrane thin, perforated everywhere with pores which lie in the meshes of the dermal skeletal reticulum. Actual diameter of pores in preserved specimens varying, often 20to soy. Dermal surface shows to the eye or lens the outer ends of small cylindrical afferent canals which pass radially inward; these appear as circular areas, mostly 175 to 300 » in diameter, abundantly scattered over the surface and covered in by the thin dermal membrane. The area of membrane covering in such a canal usually shows about three pores. Cscula small, 1 millimeter and less in diameter, scattered without order or arranged more or less definitely in longitudinal rows, which may be biseriate, viz, two rows on each branch opposite one another. In the larger chaliniform specimen (Pl. LX, fig. 21) only one branch shows anything of this regularity in the location of oscula. On this branch a biseriate arrangement appears, but it is vague; that is, irregular. In another specimen of the chaliniform type and in one of the clavate-spatulate type the oscula are arranged in short, irregular, longitudinal rows, but a biseriate arrangement is not present. In the specimen with knotty branches (Pl. LX, fig. 20) the biseriate arrangement is distinctly developed on almost all the branches, the oscula of a row lying 1 to 5 millimeters apart. Embedded in the tissue of one of the specimens are numerous embryos containing many imma- ture spicules. A good many sand grains and some large, broken, foreign monaxon spicules are embedded in the outer part of the body in the case of several specimens. Spicules (P1. LXVI, fig. 58a, b, c, d, e, f)—Megascleres: (x) Style, the chief megasclere, slightly curved, tapering toward base as well as toward point, generally smooth but sometimes spinulate, char- acteristically 110 torso pby 6to 10. Slendererones appear, especially in the connectives. Thespinu- SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 149 lation may be limited to the head, or small spines may be present over a part or a whole of the shaft. (2) Strongyle, of about same size as the style but much less abundant; obviously a modification of it. (3) Ectosomal tylostyle, commonly straight, slender, and cylindrical, sharp-pointed, head slightly tylote, 120 to 140@X24. Present but not at all abundant in the ectosome, where it is placed radially or obliquely, usually projecting. Microscleres: (4) Isochela with twisted axis, 9 to 11 w long; fairly abundant in the parenchyma; very abundant in spots in some specimens. When one tooth is seen in ventral view the other appears more or less in side view. Occasionally the rotation is greater, and both teeth are seen flatwise, one in ventral, one in dorsal, view. Normal chele, in which the axis is not twisted, also occur, but very rarely. The spicules are small and delicate, requiring to be studied with an immersion objective. In a foreshortened view the terminations appear as more or less circular cups on opposite sides of the apparently short axis. (5) Toxa, 20 to 60 2 long, in parenchyma, less abundant than the chela. Skeletal framework (P1. LX, figs. 22, 23).—Principal fibers and connectives are distinguishable. The former are primarily longitudinal, branching as they ascend, the branches curving out radially toward the surface. These fibers are polyspicular; the spicules arranged lengthwise or somewhat obliquely, and for the most part completely embedded in spongin. A few spicules project here aud there at right angles to the fiber, doubtless representing connectives that will develop. The longitudinal fibers are about 30 to 60 #7 thick. They include 3 to 8 lines of spicules as seen in optical longitudinal section of the fiber. Transverse sections show that the actual number of lines of spicules ranges in different specimens from about 3 toas many as 20. The radial parts of the fibers have usually 2 to 4 lines of spicules. The connectives are mostly one spicule in length, sometimes two, and at about right angles to the principal (longitudinal or radial) fibers. They include 1 to 2 rows of spicules, sometimes as many as three rows, when all or all but one row are quite slender. The spicules are well covered with spongin. Meshes often longer than wide, rectangular; or squarish. While the style is the chief skeletal spicule, the strongyle is very common in the connectives and perhaps predominates in them. In the ectosomal region, including a thickness of about 350 u, the principal fibers are somewhat closer together than in the interior, and there are more connectives (Pl. LX, fig. 22). The skeleton is thus denser in this region. ‘The radial fibers project slightly beyond the surface of the sponge, the terminal spicules diverging and projecting beyond the spongin of the fiber. The dermal membrane is supported by the most superficial connectives, which extend between the outer ends of the radial fibers. These dermal connectives have the usual character; that is, they com- monly have the length of one spicule, sometimes of two; the included spicules, one or two rows, are entirely embedded in the abundant spongin; common thickness of connective, 12 to 16 4; mesh squarish or polygonal. Here and there a spicule with its base rooted in the connective projects, at right angles to the latter, beyond the surface of the sponge. There is considerable quantitative variation both in the skeletal framework and the megascleres. Thus, in the same specimen the framework is somewhat denser in the older than in the uppermost part of a branch, owing largely to the fact that the connectives are more numerous and, perhaps, somewhat thicker. The differences between the several specimens in respect to these points are noticeable, although vague. In those of the chaliniform type (Pl. LX, fig. 21) the principal fibers are slenderer and the skeletal styles perhaps slenderer than in the other specimens. ‘Thus, in one of the chaliniform specimens the range in the actual number of rows of spicules contained in the longitudinal fibers is about 3 to 8, whereas in the biseriate specimen (Pl. LX, fig. 20) it is 4 to 20. In one of the clavate- spatulate specimens the skeletal style is very often noticeably stout and fusiform. But these stout styles are accompanied by a great many quite slender ones. The Beaufort sponge is not far from Esperiopsis anomala R. and D. (Ridley and Dendy, 1887, p. 84), a ramose sponge from Honolulu, in which there is a rich develop- ment of spongin, producing a chalinine appearance. Ridley and Dendy remark on this fact that it “forms a very good instance of the manner in which horny fiber may be developed in any genus.’’ Other species of Esperiopsis in which horny fiber is exten- sively developed are E. symmetrica R. and D. from off Port Jackson (Ridley and Dendy, 1887, p. 79), E. (Amphilectus) hispidula (Ridley) from Torres. Strait (Ridley, 1884, 150 BULLETIN OF THE BUREAU OF FISHERIES. p. 429), E. rigida Lambe from the Pacific coast of Canada (Lambe, 1893, p. 68). In many species of the genus the fibers consist chiefly of spicules, with comparatively little spongin. The parallelism in habitus and skeletal framework between species of Esperiopsis, Homeeodictya, and Pachychalina is noteworthy. Lundbeck has remarked (1905, p. 122) on the close parallelism between Homeodictya palmata Johnson and a species of Pachy- chalina. ‘The parallelism is equally close with E. obliqua. A similarity of another kind, involving the fundamental matter of spicule combina- tion, is presented by Esperiopsis species in general to Artemesina Vosm. This similarity is perhaps not a case of parallelism, but one due to close kinship. And in this con- nection it may be recalled that Topsent (1904, p. 215) described a species of Artemesina in which the texture of the body differs notably from that of A. suberttoides Vosmaer, etc., approaching that of Esperiopsis. Topsent suggests that it might be well to make Artemesina a subgenus of Esperiopsis, to include forms in which the body has a texture like that of Suberites. The peculiar isochelz of the Beaufort species deserve a word. They look quite like those of Microciona acerato-obtusa Carter, as drawn by Hentschel (1911, p. 349). Very small isochele, but not twisted, are recorded by Dendy (1895, p. 18) for Esperiopsis turbo (Carter); Dendy says they are very minute and difficult to detect. The occurrence of twisted chele (aniso-and iso-chele) is regarded by Vosmaer as evidence, over and above the embryological, in favor of the idea that the chela is derived from the sigma (1902, p. 9). This is at least defensible, for the twisting of the chela is in itself a structural feature that is sigmalike; that is, the twisted chela is intermediate, in respect to the shape of the spicular axis, between the sigma and the normal chela, although in other respects it is a perfectly differentiated chela. It, then, in some small measure, controverts Hentschel’s position that there are no intermediate forms between chela and sigma (1914, p. 158). Nevertheless, the spicule is ‘‘intermediate” in respect to a single point only, and this makes it very doubtful whether the point (of resem- blance) is really to be looked on as a case of reversion. It is perhaps a quite new acquisition, which happens to coincide with a phylogenetically older state. Lundbeck (1905, p. 6) thinks ‘‘the fact that chelze may be contort, a feature that is much more frequent, and may take place to a much higher degree than seems to be known by the authors, proves nothing at all” in respect to the phylogeny of the chela. Lissodendoryx Topsent (emend. Lundbeck 1905). Skeletal framework reticular, including sometimes well-marked fibers, or dendritic; spongin present more or less abundantly. Skeletal megascleres generally smooth styles, but sometimes spined; dermal megascleres diactinal. Microscleres isochele, never ancore, to which sigmata may be added. Lissodendoryx carolinensis Wilson. (Pl. LXI, figs. 26, 27, 28; Pl. LXVI, fig. 62a, b, c, d, e.) Lissodendoryx carolinensis Wilson, 1911, p. 11. Common in the harbor, especially on the wharf piles; best collecting places, Gallant’s Point, oyster cannery in Newport River, Morehead pier. The sponge begins as an incrustation on shells, ete. As it grows ascending lobes, frequently over- lapping, develop. Eventually a large amorphous mass may be produced, the body of which has been formed by the continued fusion of such lobes. The free surfaces of such masses continue to bear pro- SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 151 jecting lobes like those of the younger stages. A characteristic specimen (Pl. LXI, fig. 26) measured roo millimeters in its greatest diameter. Masses with diameter twice as great occur. Color white, frequently with a green or blue cast. Sponge comparatively firm and brittle, and generally dirty. It is much infested with worm tubes and overgrown with hydroids and polyzoa. Over the entire surface are numerous tubular translucent papillae (Pl. LXI, fig. 27), perforated by numerous pores. The papilla may be simple or branched, often bifurcated. They are contractile and may almost entirely disappear. Oscula 1 to 2 millimeters in diameter are scattered over the surface and often develop at or near the ends of lobes. The pores are distributed over the general dermal membrane and papille. Over the papillz they are abundant and in an alcoholic specimen measure about 20 » in diameter. In the same specimen the pores over the general surface are almost all closed. The few found open measure up to 100 w in diameter. The dermal membrane is translucent, showing anastomosing subdermal canals, commonly about 0.5 to 0.75 millimeter wide. Spicules (Pl. LXVI, fig. 62a, b, c, d, e).—Megascleres: (1) Style, smooth and slightly curved. Range of size, 160 to 180 w by 5 to 8 wu. (2) Tylote, smooth. Range of size, 160 to 190 w by 5h. Microscleres: (3) Isochelz arcuate 12 to 264 long. (4) Sigmata 18 to 36 u long. Skeletal framework (Pl. LXVI, figs. 27, 28).—Internal skeletal framework is a loose, irregular reticulum formed by styles, which may in places develop into spiculo-fibers. Meshes of reticulum are three to five sided; side of mesh about the length of a spicule, formed by one, two, or three spicules. Spongin present at the nodes (stained sections show it). In addition to the skeletal reticulum, the parenchyma contains some scattered tylotes. These may be grouped to form loose tracts. The tylotes are especially abundant in collenchymatous regions, and are more abundant in the ectosome than elsewhere. The megascleres of the dermal membrane are tylotes. In places they are scattered tangentially in the membrane, but very generally they project more or less radially, forming bunches or ridges. The wall of the papilla (Pl. LXI, fig. 26) is an extension of the dermal membrane, and the megascleres here, too, are tylotes, tangentially arranged and forming a reticulum, in the meshes of which are the pores. The whole parenchyma is loaded with sigmata; isochele abundant, but less abundant than the sigmata. Both sigmata and isochele are abundant in the general dermal membrane. ‘The wall of the papillz contains moderately abundant isochele and very few sigmata. Topsent established Lissodendoryx first as a subgenus of Dendoryx (1892) and later (1894a, p. 9) as a separate genus, for species which differ from Dendoryx (=Myxilla sens. str., Thiele, 1903; Lundbeck, 1905; Topsent, 1913) in having smooth styles as the skeletal megascleres. Dendy (1895, p. 29) would include the genus, and Dendoryx as well, under Myxilla, Schmidt. Topsent (1901), p. 19; 1904, p. 173) retains the genus as originally defined. Lundbeck (1905, p. 153) again brings up Dendy’s contention (1895) that the smooth- ness of the styles can not be used as a generic character, since species occur with styles that are intermediate between spined and smooth ones. From this point of view the genus should be merged in Dendoryx (= Myzxilla sens. stv.). But the Dendoryx species, Lundbeck maintains, are separable into two groups, in one of which the microscleres are ancore, in the other chelz arcuate. For the former Lundbeck reserves the name of Myxilla (sens. str., Topsent, 1913, p. 623), for the latter Lissodendoryx. 152 BULLETIN OF THE BUREAU OF FISHERIES. Subfamily PHL@ODICTYINA. Sponge body provided with fistular outgrowths. Characteristically the ectosomal skeleton is much denser than the choanosomal, constituting a sort of rind. The micro- scleres are often absent. Phloeodictyon Carter. Spongin usually present, but the skeleton is not a reticulum of distinctly chalinine spiculo-fiber. Megascleres, oxeas varying to strongyles. There are no microscleres. Phleeodictyon nodosum n. sp. (Pl. LXII, figs. 29, 30, 32; Pl. LXVI, fig. 63.) One specimen dredged in Beaufort Harbor. Slender fistulz, 30 to 35 millimeters high and 1.5 to 3 millimeters in diameter, rounded off and closed terminally, connect with a basal portion which is attached to a piece of shell. The basal portion isincomplete. As it stands it consists of an incrusting part on the upper surface and a somewhat thicker, 2 to 3 millimeters thick, torn part on the under surface of the shell. The entire sponge was probably not large. The shell is probably to be looked on as having been surrounded by, and incorporated in, the upper part of the sponge body. Color, in alcohol, whitish brown. Wall of fistule thin, but firm. Many sand grains and pieces of shell have been incorporated by the sponge. The dermal membrane of the fistular wall is perforated by pores lying in the meshes of the skeletal reticulum. Many of the pores are closed, and those that are not closed are probably only partly open. They measure 12 to 16 w in diameter. The membrane is thin, contains only a few granular amcebocytes, and is favorable for histological study. Some few foreign incrustations cling to it, among them holothurian (synaptid) spicules such as Bowerbank has figured (1864, Pl. V, figs. r19, 120). Pores and oscula over basal part of sponge uncertain. This part exhibits a good many small canals, 1zoto 5scouindiameter. Flagellated chambers spheroidal, or about so, 28 w in diameter. The ectosome of the fistular wall includes a great many small, rounded subdermal cavities, about 60 to roo u in diam- eter as seen in cross sections of the fistula (PI. LXII, fig. 30). Internal to the ectosomal skeleton, the fistula is collenchymatous and is excavated by a large axial canal around which lie smaller canals, which yet are of good size, about 150 to 350 m in diameter (fig. 30). These are separated by thin sheets and strands of sponge tissue. All the fistule are closed terminally, showing no sign of oscula. If ordinary oscula were present in life, one would expect to see some sign of them in the preserved specimen. Perhaps the axial canal opens terminally through a sieve plate, the apertures of which, resembling pores, are now closed. As Lundbeck says (1902, p. 58), there is diversity of opinion with regard to the functioning of the fistule in these sponges. Living, or, at any rate, carefully preserved whole specimens need to be studied. The pores over the general surface and the subdermal cavities of the ectosome make it clear that water streams into the fistula. Nevertheless, perhaps the axial canal is efferent. It would seem that it must be so in species such as P. elongatum Tops., where it connects with the exterior by a terminal or sub- terminal aperture which has the appearance of being normal (Lundbeck, loc. cit., p. 60): Spicules.—Oxeas (Pl. LXVI, fig. 63) smooth, slender, slightly curved, subcylindrical, tapering gradually to sharp points, about r00 nz by 4to 54. The strongylate modification occurs. Skeletal framework of the fistule.—The ectosomal skeleton includes the usual parts, a dermal layer of tangential spicules and a subjacent layer of spiculo-fiber. The dermal spicules intercross in all directions, constituting a layer which is, in general, single, although in places parts of two or three spicules may be superposed. The spicules form a reticulum (Pl. LXII, fig. 32), at the nodal points of which they are heid in place by spongin. These nodal points in astained preparation of the wall are conspicuous. Meshes of the reticulum triangular or polygonal; side of a mesh the length of aspicule or less and formed by one, or, in places where the spicules are more abundant, by several, about two to five, spicules. In regions of the latter character the spicules are so closely grouped that they radiate from many of the nodal points, like spokes of a wheel. Elsewhere in the same fistula the reticulum may be unispicular. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 153 Beneath the dermal layer of the fistula and about a spicule’s length from it lies the fibrous layer. The two are connected by smali bundles of spicules or by single spicules, which extend radially or obliquely to the surface between the subdermal spaces (fig. 30). The fibrous layer consists of polyspicular spiculo-fibers (figs. 30 and 32), which pursue a longitu- dinal course, dividing and anastomosing to some extent, interconnected by slender tracts of spicules or by scattered spicules forming a secondary reticulum. The secondary reticulum varies in character in the same fistula. The mesh may be polygonal, with a side equal to the length of a spicule, unispicular in one region, polyspicular in another; or the spicules may be thickly and confusedly crossed, giving the mesh a side less than the length of a spicule; or the mesh may still be polygonal, but its sides two or three times the length of a spicule and formed by polyspicular tracts. At the nodal points of the reticulum there isspongin. The main longitudinal fibers are 30 to 80 u thick; spicules of a fiber arranged lengthwise, closely packed, and bound together by a very small amount of spongin, which does not form a covering for the fiber. Internal to the fibrous layer the fistular wall contains almost no skeleton. At most a few spicules project radially and obliquely inward from the fibrous layer, and here and there a free spicule is found. Skeletal framework of basal part of sponge.—Where the surface of this part of the sponge has been preserved, an ectosomal skeleton is found much like, nevertheless somewhat different from, that of the fistula. There is a dermal layer, two or three spicules thick, of tangential megascleres, which cross in every direction. These are united, as in the fistular wall, by spongin at the nodal points, but the spicules are so abundant that they can not be said to form a reticulum. Still they are not as densely packed as in some species; everywhere minute angular spaces, commonly about 12min diameter, are left be- tween them. Beneath the dermal layer and visible through it spiculo-fibers, like those of the fistula, form a coarse and very irregular reticulum. The choanosomal skeleton consists chiefly of abundant scattered single spicules, crossing one another in all directions, without spongin. Here and there the spicules are combined to form loose tracts. The Beaufort species is nearest P. reticulatum Tops. from the Azores (1904, p. 238). But in the latter the spicules of the dermal layer are loosely intercrossed; the fibrous layer of the ectosome is a network with subequal, round, or oval meshes; and the oxeas are larger, 175 to 210 uw by 3 to 13 yn. Lundbeck (1902, p. 56) dissolves Carter’s group Phloeodictyine and distributes the genera (Rhizochalina, Phlceodictyon, and Oceanapia) among the Chalinine, Renier- ine, and Geiliine. The treatment is logical if we regard only the spicules (and, in the case of Rhizochalina, the character of the spiculo-fiber). Topsent (1904) and others have followed Lundbeck. Dendy (1905), keeping in mind the presence of fistule and the dense ectosomal skeleton, constituting a rind, retains the group, adding to it Histoderma, Sideroderma, and Amphiastrella, which necessitates placing it in the Desmacidonide. The group, as Dendy remarks, seems to be a natural one. Subfamily PHORIOSPONGINZ: Lendenfeld, 1888, 1889, ernend. The skeletal fibers are very areniferous, sometimes partly spicular; they may be reduced to rows of sand grains united or not by spongin. Skeleton usually reticulate, but sometimes consisting of independent fibers or of scattered sand grains. The mega- scleres are monaxonid, monactinal or diactinal, or both. ‘The microscleres are chelze and sigmata, but either or both may be lacking. The flagellated chambers are (always?) eurypylous and large. 154 BULLETIN OF THE BUREAU OF FISHERIES. Phoriospongia Marshall, emend. [With the characters of the subfamily.] Phoriospongia, Marshall, 1880. Chondropsis (Sigmatella Lendenfeld), Dendy, 1894. Psammochela, Dendy, 1916. Phoriospongia osburnensis, n. sp. (PI. LXI, figs. 24, 25; Pl. LXVI, fig. 60a, 5, c.) A single specimen (Pl. LXI, fig. 24) taken on the “Fishing Bank’’ off Beaufort Inlet by Dr. R. C. Osburn at a depth of 13 fathoms. Sponge forms a thin incrustation over an alcyonarian coral. It is for the most part about 1 milli- meter thick, thinner in places, and twice that thickness in some spots. Color whitish in alcohol, on the salmon-pink alcyonarian. Oral ends of the alcyonarian polyps in general free of the incrustation. Loxosoma is scattered in abundance over the surface of the sponge. Surface fairly smooth. Pores 50 to 80 in diameter, abundantly scattered over the dermal mem- brane, which is quite riddled with them in many places, probably everywhere when they are open. Small, rounded subdermal cavities, mostly 125 to 250 wide, are very abundant and give to the sur- face of the sponge, when examined with a lens, a porous appearance. Oscula uncertain; probably small and scattered, and now closed. Many canals, the largest about 200u wide, excavate the paren- chyma, some passing radially through the incrustation from the surface almost to the base. Flagellated chambers uncertain. Abundant, small, compact cellular masses, doubtless embryos, occur in the paren- chyma. Spicules (P1. LXVI, fig. 60a, b,c). —Megascleres: (1) Strongyles, subcylindrical with bluntly rounded ends, slender, smooth; about straight or slightly curved; 160 to 180% by 2 to 34. Microscleres: (2) Sigmata, 10 to 20 uw long; the common and characteristic length, 14 to 164. Abundant in dermal mem- brane; scantily present in the interior, canal walls, and parenchyma. (3) Isochele, 12 to 16 long; very scantily present in dermal membrane and interior, The axis is distinctly curved and the spicule is tridentate, the toothed end about one-fourth the total length. The teeth appear, under a good immersion objective, elongate-conical, but the small size of the spicule makes minute details some- what uncertain. The spicule probably falls in the Levinsen-Lundbeck (Lundbeck, 1905, p. 2) class of chele arcuatz; that is, there is at each end only one tooth proper, the lateral “‘teeth’’ being the ale, which are separated by deep bays from the shaft. These spicules were at first overlooked in the sponge, but after their discovery search revealed a few in every preparation. Skeletal framework (P1. LXI, fig. 25).—There is no reticulum. Instead, simple, unbranched, skeletal fibers pass more or less radially from the base of the sponge to the surface (fig. 25). They often curve a good deal, and the precise direction of the fiber frequently corresponds with that of an adjacent large canal; obviously a case of correlation. The fibers are made up of sand grains, with which some bits of foraminifer shells or pieces of foreign sponge spicules are intermingled, proper spicules (strongyles) of the sponge, and spongin. The sand grains of the fiber and bits of shell, etc., are frequently but not always arranged in a single series. The grains may be all small, or here and there a much larger one is intercalated. Covering them sparingly are strongyles, arranged more or less lengthwise in the fiber. There are short parts of many fibers in which no sand grains are present; these are composed of compact tracts of longitudinally placed strongyles. The spongin of the fiber is scanty. Yet there is enough not only to connect but to form a thin covering over the mineral elements of the fiber. It is very transparent and easily overlooked. The skeletal fibers in the ectosomal region are frequently made up chiefly of proper spicules, stron- gyles, which at the surface project slightly, often diverging. Beside these radial or obliquely radial bunches, free megascleres occur here and there in the ectosome, inclined more or less radially to the surface. The skeletal fibers, owing to their composition, vary greatly in thickness, even within the same fiber. Parts of fibers may be only 20 thick, other parts 160 thick. The fibers are abundant, frequently 175 to 350 u apart. Between the skeletal fibers there are, in the body of the sponge, some scattered megascleres. The dermal membrane contains a good many strongyles lying tangentially, scattered singly or in wisps; also the projecting tufts of spicules, above referred to, which represent the upper ends of skeletal fibers. Flat preparations seem sometimes to show skeletal fibers running tangentially in the dermal membrane. But sections prove there are no such fibers, and that the appearance is caused by fibers which run from base to surface as usual, but very obliquely and in regions where the sponge is quite thin. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 155 We had at first decided to inscribe this sponge under Dendy’s recent genus Psam- mochela, the diagnosis (Dendy, 1916, p. 126) of which reads: “‘Reticulate skeleton composed of sandy and sometimes partly spicular fibers. Magascleres styli or strongyla, or both. Microscleres isochele, which may be very minute and with vestigial teeth; to which sigmata may be added.”’ But we now feel, for the following reasons, that the subdivision of the Phoriospongine into Phoriospongia, Chondropsis, and Psammochela is not satisfactory. The distinction between Phoriospongia and Chondropsis (Sigmatella) is arbitrary. Lendenfeld (1888, 1889) based the distinction especially on the sigmata. Forms with large sigmata, 30 to 504 long, were put in Phoriospongia, those with very small sigmata, 5 to row long (Lendenfeld’s figure indicates that the 1 » given in the text as the length of the sigma in Chondropsis australis is a misprint for 10 w. Vide Lendenfeld, 1889, p. 611.), or with none, in Chondropsis (Sigmatella). Dendy (1894, p. 250) found it necessary to change Sigmatella (preoccupied) to Chondropsis, and further pointed out that the size of the sigmata could not justly be used as a mark by which to distribute the species of the subfamily. In recorded species the sigmata measure 54, 10M, 164, 30m, 35M, 50 in length, thus forming a fairly con- tinuous series. Dendy nevertheless retains the two generic names Phoriospongia and Chondropsis, and would assign to the former species with monactinal megascleres, and to the latter those with diactinal megascleres. Hence, several forms listed by Lendenfeld under Phoriospongia are shifted by Dendy to Chondropsis. But Dendy’s basis for the distinction between the two genera can not be thought of as satisfactory since the character of the magascleres is variable in these sponges, as is borne out by the following: Lendenfeld (1889) records that the megascleres in Chondropsis (Sigmatella) australis are chiefly strongyles but in part styles and tylotes; in C. turbo they are strongyles with some styles; in C. corticata strongyles but also in part oxeas and styles. In Dendy’s new addition to the subfamily, Psammochela (Dendy, 1916, p. 126), the megascleres are styles or strongyles or both. It seems therefore necessary to merge Chondropsis into Phoriospongia. As to Psammochela, its distinction from Phoriospongia rests on the presence of chele. It does not seem justifiable, however, to separate from Phoriospongia forms like the Beaufort species in which the chele are so scarce as to be easily overlooked. Rather we may conclude with a good deal of probability that actual search will reveal a scanty number of chelz in some, at any rate, of the forms hitherto listed under Phorio- spongia and Chondropsis and supposed to be without these spicules. Further, it may be recalled that in one of the specimens of Ridley’s Phoriospongia fibrosa (Ridley, 1884, P- 439) chele were found to be scarce, in the other abundant. This indicates that it is not rational to separate the forms with abundant chele from those with few or none. Phoriospongia should therefore be emended to include forms both with monactinal and diactinal megascleres, and those in which chelz persist either abundantly or in small number. It thus becomes coextensive with the subfamily, and some artificial grouping of the species may be desirable as facilitating reference to them. As to the position of the genus, Lendenfeld (1888, 1889) made his Phoriosponginz a subfamily of the Spongelide. Dendy (1894, p. 250) and Topsent (1894), p. 5) trans- ferred the genera to the Monaxonida, placing them near their supposedly closest rela- tives, the Gellius-like sponges. Dendy more recently (1916, p. 126) utilizing the data 156 BULLETIN OF THE BUREAU OF FISHERIES. afforded by his P. (Psammochela) elegans, has shown that the sponges belong in the Desmacidonide, although the chele have apparently been lost by many species. The species of this subfamily, except the Beaufort form, are all known to reach a large or, at any rate, a good sizer Tubular and flabelliform shapes and a massive or irregular form of body with processes are common. At least three species are known in an incrusting phase, P. (Sigmatella) corticata papillosa (Lendenfeld, 1889, p. 620), P. (Sigmatella) carcinophila on crabs (Lendenfeld, loc cit., p. 615), and P. (Desmacidon) psammodes (Hentschel, 1911, p. 322; Dendy, 1916, p. 126). Perhaps the Beaufort species will be found in some larger phase. ‘That it breeds in the thin, incrusting con- dition is no reason for believing that this is its final state. Mucrociona prolifera, for example, breeds in Beaufort Harbor, while a thin incrustation having a skeletal arrange- ment much like that of P: osburnensis (Wilson, 1911, Pl. I, fig. 5); but, while the great majority of individuals in the harbor do not get beyond an incrusting condition, much larger and more complicated phases are reached by some. If P. osburnensis reaches a large size, probably its skeleton becomes reticulate. At present it is the only recorded form, except P. solida Marshall, which lacks recognizable fibers, in which the skeleton is not reticulate, although in P. (Sigmatella) carcinophila Lendenfeld (1889, p. 615) the reticulum is confined to the basal portion of the sponge, while from it isolated vertical fibers pass to the surface. The complete loss of spongin in some forms, if it really occurs, is a remarkable fact, especially since the skeletal sand grains continue to be arranged by the sponge in bands. Perhaps the spongin is really not absent in these forms but only very scanty and transparent. Ridley (1884, p. 439) says that in P. fibrosa the skeletal “fibers are wholly composed of foreign bodies united by an almost colorless, not dense, substance,” and, as stated above, the spongin in P. osburnensis is easily overlooked. The species of the subfamily hitherto recorded are all from Australian waters. ‘Two of the forms are, however, thought by Lendenfeld (1889, pp. 613, 620) to occur elsewhere, P. (Sigmatella) australis var. tubaria at Nassau (Bahamas) and P. (Sigmatella) corticata var. papillosa on the English coast, African coast, in the Indian Ocean, and on the Florida coast. The last-named item of distribution rests on the identification of Hyatt’s Spongelia kirkii (Hyatt, 1877, p. 539) as a Phoriospongia (Sigmatella), but I can not find that Hyatt’s account justifies this step. The chele in P. osburnensis are described (see above) as ‘‘tridentate.’”’? Dendy describes (1916) in the same way the chele of his P. (Psammochela) elegans, the smaller of which (fig. 6c, c’) resembles that of P. osburnensis; and in P. jfibrosa Ridley (Ridley, loc cit.) the isochele are said to be ‘‘tridentate.” WLundbeck (1905, p. 4) criticizes the use of this term (category) as confusing, since it covers two different forms of spicule, chele arcuate and ancore. ‘This is undoubtedly so. On the other hand, it must be allowed that when the cheloid is very small, it is difficult to use the Levinsen-Lundbeck categories. For, assuming that the “tridentate” spicule is really either a well-defined chela arcuata or ancora with three teeth, and not some other form, the decision turns on whether the shaft has, in addition to the three teeth, ale or not. And this is not easy to determine with certainty when the spicule is very small. Hence it would seem allowable, even necessary, to continue to use in practice the term “tridentate” for certain small spicules, although it is confessedly somewhat vague. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 157 Levinsen and Lundbeck (Lundbeck, 1905, p. 6) regard the distinction between chela and ancora as so fundamental that they use it in distributing the species into genera, species with chele being assigned to one genus, species with ancore to another. Dendy apparently does not support this practice, for he includes in the same genus (1916, p. 126) a form, Psammochela elegans, with isochele (they are so designated, and the figures show three palmate teeth and no additional ale) and one, Desmacidon psammodes Hent- schel, with ancore. Hentschel, on the other hand, uses the distinction, at any rate in some cases, that of Desmacidon-Homeodictya, e. g., (Hentschel, 1911). Considering the existence of small cheloids, it seems to us that the distinction is one which can not be rigidly used in distinguishing genera. Subfamily ECTYONINZ. Skeletal fibers, or spicular tracts, with echinating spicules which are character- istically spinose (acanthostyles). Microciona Bowerbank. Sponge body incrusting, or the incrustation may develop lobes and with continued growth become a complex, branched, ascending mass. Skeleton originally a basal plate bearing short, upright, plumose columns. In older forms the skeleton becomes an internal reticulum of spiculo-fiber, beset with short, plumose, radial fibers which represent the upright columns of the incrusting phase. Megascleres monactinal; the chief spicules smooth styles, the echinating spicules smaller and more or less spinose. Microscleres isochele, often accompanied by toxas and sometimes by sigmata. Microciona prolifera Verrill. (Pl. LXII, figs. 31, 33; Pl. LXIII, figs. 35, 36; Pl. LXVI, fig. 57a, b, c, d, e.) Microciona prolifera Verrill and Smith, 1874, p. 447. Microciona prolifera Wilson, 1902, p. 396. Microciona prolifera Wilson, 1911, p. 3. The sponge when young forms thin incrustations on oyster shells, wharf piles, etc. As it grows older there rise up crooked, irregular lobes (Pl. LXIII, fig. 36). Asthe sponge grows older, the growth and formation of lobes may continue. This continued growth accompanied by branching and anas- tomosis will ultimately produce an intricately branched sponge (Pl. LXIII, fig. 35). Specimens of this type may reach a height of 150 millimeters. The oscula are small apertures scattered here and there over the surface in general. They lead into canals which extend tangentially just beneath the dermal membrane. ‘The pores are irregularly scattered in considerable abundance over the surface and lead into subdermal spaces. Thus the cavi- ties which immediately underlie the dermal membrane are of two kinds, some afferent and some effer- ent. In the lobes of young specimens and in older branched specimens these superficial spaces com- municate with abundant canals which ramify throughout the sponge interior (PI. LXII, figs. 31 and 33). Spicules—Megascleres: (1) Styles (Pl. LXVI, fig. 57a, 6), smooth and slightly curved, measuring 150 to soon by 8 to 12 in an incrustation; 150 to 3804 by 8 to 14m in one of the lobes of a young specimen; and 150 to 380u by 8 to 16uin an older branched sponge. The styles frequently have slightly enlarged heads and sometimes the heads are beset with very minute spines. (2) Small spinose styles 80to toon by 6 to 8u, which frequently have slightly enlarged heads (fig. 57c). Large numbers of very slender young megascleres are found throughout the sponge. Microscleres: (3) Isochele, 12 to 16p long (fig. 57d). (4) Toxas, 10 to 4op long (fig. 57e). Wilson (loc. cit.) records the microscleres, espe- cially the toxas, as scanty in Beaufort specimens. We find that the granular sponge tissue tends to obscure the microscleres and that in partially macerated sections they may be found scattered in con- siderable abundance. 110307°—21——11 158 BULLETIN OF THE BUREAU OF FISHERIES. Skeletal framework.—The skeleton of the incrustations consists of a horny basal plate bearing closely set vertical horny columns from which megascleres project. From near the apex of each horny column a few large, smooth, and slightly curved styles project, forming a well-marked tuft. These styles measure 160 to 4oou by 8 to rou. The longest styles lie near the apex of the column, and some of them project beyond the surface of the sponge. Mingled with the mature styles are younger spicules of the same type, but slenderer and shorter. Projecting from the sides of some of the larger horny columns are a few small styles, 80” by 5 to 6u, some of them distinctly spinose, others with few and feeble spinulations. The skeleton of the incrustations contains longer styles than are found in the lobes of specimens like figure 36 or in the branches of older specimens like figure 35. In the incrustations we found a good many styles measuring 5oou long, while in the lobes and branches of older specimens they rarely exceed 380 in length. n The skeleton of the constituent branches in a specimen like figure 35 consists of a reticulum of horny spiculo-fiber (Pl. L-XII, figs. 31 and 33) which breaks up near the surface into independent radial fibers that extend out to and support the dermal membrane. From near the apex of such a radial fiber a few large, smooth, slightly curved styles project, forming a well-marked tuft. The longest styles are found near the surface, and many of them project beyond the surface of the sponge. These large styles average about 3304 by rou. Projecting vertically and obliquely from the sides of the radial fibers are large, smooth styles similar to those near the apex and a few small distinctly spinose styles (about Sou by 7), together with others of about the same size but with few spinulations. It is these small styles which represent the echinating spicules of the subfamily. The spiculo-fibers of the inte- rior bear similar echinating spicules. Wilson has pointed out that “‘the projecting (echinating) styles are few and scattered, spinose or smooth, the two types intergrading. The spinose type has numerous distinct though small spinulations on the shaft and a minutely tuberculate head. Spicules with only a few scattered spines occur, and, finally, quite smooth spicules with head end simply rounded and not enlarged.” The description given above applies both to the lobes of young specimens and the constituent branches of older ones. The skeleton of the older specimens (Pl. L-XIII, fig. 35) differs, however, from that of the younger (fig. 36) in the following details. The spiculo-fibers in the former are consid- erably thicker than in the latter, due to the increased accumulation of spongin, and the styles reach a greater thickness. Some styles in the older sponge were found to measure 16uin diameter while the greatest thickness observed in the younger sponge was 14u. The small echinating styles are also more abundant in the former. Family AXINELLIDA. Sponge body ordinarily more or less upright, of a branching, lamellate, or cuplike habitus. But massive and even incrusting forms occur. Skeleton typically consists of ascending bundles of spiculo-fibers, from which arise subsidiary fibers that radiate to the surface. Skeletal fibers without spined echinating spicules, and typically plumose. The characteristic megascleres are monactinal. In addition to these, diactinal mega- scleres may also occur, and in some genera are the only form. Microscleres in the shape of microxeas, trichodragmata, or sigmata occur in a few genera; cheloid microscleres do not occur. Axinella O. Schmidt. Sponge body typically ramose; habitus varies, however, but while sometimes lamellate it is noc cuplike. There is a firmer axial skeleton from which radial fibers pass to the surface. Axial skeleton not massive, but made up of ramifying and anasto- mosing spiculo-fibers. The radial fibers, which terminate in brushes of diverging smaller spicules, are joined by numerous short transverse connectives, the genus differing in this point from Phakellia in which the radial fibers are comparatively free. Skeletal fiber more or less distinctly plumose. Megascleres chiefly styles, but strongyles and oxeas may occur; scattered acanthostyles may also occur sparsely, as a vestigial feature. Microscleres generally absent, but trichodragmata are present in some species. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 159 Axinella acanthifera, n. sp. (Pl. LXII, fig. 34; Pl. LXIII, figs. 38, 39; P!. LXVI, fig. 50a, b, c, d.) One specimen taken on Fort Macon beach. Sponge body (Pl. LXII, fig. 34) lamellate, narrowing below to a stalk, expanded and divided into lobes above. Lobes in general foliaceous, separated by marginal notches; but the margin is also pro- duced here and there into short, subcylindrical lobes. Upper part of sponge curved irregularly, so that the lobes lie in different planes. In this particular specimen growth has evidently resulted in an early division of the main axis, the two parts thus produced later on fusing at a point higher up. Height of specimen, 40 millimeters; width, 30 millimeters; thickness of lamella, 2 to 3 millimeters. Color in alcohol gray brown with a tinge of yellow. Sponge firm, but flexible. Surface smooth and velvety. Both surfaces are alike, appearing finely porous to the eye. The surface is depressed between the dermal brushes of projecting spicules; where greatly depressed, this is probably due to drying. Pores 60 to 85 » in diameter are abundantly scattered between the dermal brushes. Fine canals less than o.5 millimeter in diameter may be seen here and there beneath the dermal membrane, and parallel to the surface of the sponge. In places they radiate toward a central stripe, in which two or three minute apertures, about o.5 millimeter in diameter, are arranged in a row. ‘These are doubtless oscula. A very similar arrangement is described for Axinella manus Dendy, from the Gulf of Manaar. “The vents are small openings in the floors of stellately arranged or longitudinal grooves’’ (Dendy, 1905, p- 189). Spicules.—(1) The most abundant megasclere is a smooth, slightly curved style, 160 to 260 u long by 7 to 12 w thick; commonly about 210 by 8u (Pl. LXVI, fig. 59a). This spicule makes up the bulk of both axial and peripheral skeleton. In the dermal brushes these styles, which constitute most of the brush, are thicker than elsewhere; common range, 160 to 200 u by 1o to 12 w. (2) A much stouter style, straight or nearly so, 160 to 240 uw by 12 to 20 u (fig. 59) is intermingled with the common form both in the axial and peripheral skeleton. (3) A very long and slender style, 400 to 600 u by 6 to 7 uw (fig. soc; Pl. LXIII, fig. 38, right side), is a characteristic element of the dermal brushes. Each brush includes a few (one to four) of these spicules, which project far beyond the others. They are broken off over much of the surface. (4) A small spinose style, 80 to 120 u by 6 uw, with strong spines (fig. sod), is present in the dermal brushes, in the radial fibers, in and projecting from the connectives that ex- tend between the radial fibers. The spicule is not common anywhere, but is easily found on searching. Only those few which project from the connectives could be classified as “echinating’’ spicules. The bulk of them occupy positions similar to those of the common style. (5) Long and very slender rhaphid- like spicules occur in considerable abundance, scattered singly or in loose irregular sheaves; charac- teristic spicules measure 200 by 1 4; more abundant in the ectosome than elsewhere. ‘The sheaves are not to be confused with trichodragmata. It is questionable whether these spicules are rhaphides or simply stages in the development of the megascleres. Rhaphides in bundles have recently been des scribed in a related sponge, Raspailia (Syringella) rhaphidophora Hentschel from the Aru Islands (Hentschel, ro12, p. 371). Skeletal framework.—The skeleton (P1. LXIII, figs. 38, 39) is divisible into an axial and a peripheral part. The axial skeleton is chiefly composed of longitudinal fibers which anastomose to some extent, but the unions between which are more commonly made by transverse or oblique connectives irregu- larly disposed. The longitudinal fibers are polyspicular, with abundant spongin, the spicules arranged longitudinally in the fiber or obliquely, with the point slightly projecting. The fibers vary greatly in thickness and character in different regions of the sponge. Inthe upper part they are about 30 to jon thick and well filled with spicules, the spaces between them wider than the fibers themselves. Nearer the base the axial reticulum is closer and more compact. The individual fibers are here thicker, owing to the increase in the amount of spongin around the spicules, the latter now forming only an axial core. The fibers in the basal region range from about 60 to 120 uw in thickness, the interstices between them being about the thickness of or narrower than the fibers. The connectives between the longitudinal fibers include, as arule, one to three spicules, which are covered with abundant spongin. The spicules of a connective are often longer than the space between ~ the fibers that are joined, and thus cross both fibers. The peripheral skeleton is made up of radial fibers, including their dermal terminations, and con- nectives. The radial fibers are prolongations of the axial fibers that curve in an obliquely radial direction to the surface. They branch to some extent. The fibers are polyspicular; the spicules held 160 BULLETIN OF THE BUREAU OF FISHERIES. together and just covered with spongin; the spicules longitudinal in the fiber or slightly oblique, with apex projecting. The fibers enlarge gradually as they approach the surface. At the outer end the spicules diverge somewhat, forming a dermal brush, which projects beyond the surface of the sponge. Connectives extend between the radial bundles at about right angles to the latter. They include each one to three spicules, covered with spongin. The dermal skeleton is made up of the projecting dermal brushes and the most superficial connec- tives. Axinella is often defined as having plumose skeletal fibers. This term applies fairly well to the outer parts of the radial fibers in the Beaufort species, but not well to the rest of the skeleton, although the axial fibers and the inner parts of the radial fibers all show plenty of spicules placed obliquely, with the points slightly projecting; and after all, it is to this position of the spicule in the fiber that the ‘‘ plumose”’ character is reducible. In its external form this sponge closely approaches the European Axinella flustra (padina) Topsent (Topsent, 1896, p. 131; 1904, p. 139), a species which has tricho- dragmata about 40 wp long. If we lay too great a stress on the presence of the spined styles, A. acanthijera might be removed to the Ectyonine and put in or close to Raspailia, in some of the species of which the acanthostyles are vestigial or absent, as in subgenus Syringella (which, to be sure, certain authors separate completely from Raspailia, making it an independent genus placed in the Axinellide). This is the familiar form of argumentation based on the presence or absence of a particular feature which leads in its application to the establishment of ‘parallel genera,” viz, genera assigned to different groups, which yet resemble one another except in regard to the feature in question. From one point of view such parallel genera are looked on as temporarily defensible because of their practical utility, although artificial. From another they are regarded as natural groups which owe their general similarity to independent adaptation (or ‘‘convergent evolu- tion”). But the character of the skeleton taken as a whole (cf. Vosmaer, 1912, p. 310) leaves little doubt that the sponge belongs with the other Axinella species. The presence of the spinose styles is accordingly to be interpreted as a vestigial feature which has never been quite lost, or, possibly, as an imperfect return (incomplete rever- sion) to an ancestral condition which had disappeared in the Axinellids. The sponge resembles Raspailia, or, at all events, the species which center around R. viminalis O. Schm. (cf. Vosmaer, 1912, p. 313), in yet another respect, viz, in the presence of long projecting styles in the dermal brushes. In the R. viminalis type, regarded by Vosmaer (loc. cit.) as characteristic of the genus, each dermal brush includes a single, strong, far-projecting style, surrounded at its base by a tuft of diverging small spicules, generally styles, sometimes oxeas. There are, of course, many species gen- erally assigned to Raspailia, in which the radial fibers (or the tufts of spicules which represent them) lack the large terminal styles, and a few such as R. irregularis Hent- schel from the Antarctic (Hentschel, 1914, p. 121) and R. hornelli Dendy from the Gulf of Manaar (1905, p. 172), in which each dermal brush includes not a single such style, but a bunch of them. It will be noticed that in the Raspailia species the long-projecting styles are stouter than the surrounding spicules of the brush, whereas in the Beaufort sponge they are slenderer. Nevertheless, their presence, coupled with the occurrence of acanthostyles in the Beaufort sponge, greatly strengthes the already well-supported view that Raspailia is an intermediate form, leading up from the Ectyonine to Axin- SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 161 ellids, such as Axinella. Any cleavage of this close series of forms, all the members of which continue to exist, into genera is largely a matter of convention based on historical accidents; that is to say, certain terms and not others of the series came to be studied first, and so became the nuclei of genera. As to the course of opinion with respect to the position of Raspailia, it may be recalled that Ridley and Dendy (1887) first gave the genus a definite postion in modern classification by assigning it to the Axinellide. Topsent (1894b) removed it to the Ectyonine. Dendy (1895, p. 46) assented, but remarked that the genus was inter- mediate between the two groups, Ectyonine and Axinellide. ‘Topsent has been fol- lowed generally, but Vosmaer (1912), without mentioning groups by name, keeps the genus in the same series with Axinella, Phakellia, Acanthella, ete. Dendy (1905, p. 172) apparently departs from his former view and now regards the resemblance of the Raspailia species to Axinellide as strong but “superficial.”’ He says “it is evident from the presence of the spined echinating styli (though these may be vestigial) that they are really highly modified Ectyonine.’’ The Axinellide are looked on as a poly- phyletic group. Dendy recently (1916, p. 96) is disposed to abandon the group entirely and to include the genera such as Axinella, Phakellia, etc., in the Haploscleride, thus virtually returning to Vosmaer’s position (1887, p. 335). Hentschel (1912, p. 413) brings out the skeletal resemblances between the Ectyonine genera centering around Raspailia and certain Axinellide, and inclines to regard it as due to kinship and not to convergent evolution (position of Dendy in 1895). Acanthella O. Schmidt. Form generally lamellate, consistency cartilaginous. Surface aculeate or conu- lose. If radial fibers are developed, they are weak, except in the basal part of the sponge, as compared with those of Phakellia; connectives between the radial fibers are lacking. Without microscleres. Acanthella corrugata, n.sp. (Pl. LXIII, fig. 37; Pl. LXV, figs. 46, 47; Pl. LXVI, figs. 56a, b, c,d, e.) One specimen taken on Fort Macon beach after a moderate southwest blow. Sponge body (Pl. LXIII, fig. 37, taken obliquely from above) afolded lamella; folds tend to anastomose and inclose cuplike compartments. Only one of the cuplike spaces is completed, and this is open at the bottom; there are several other partially surrounded spaces. Below, the sponge narrows to a short peduncle; in side view the body has about the shape of an open fan. Total height, 65 millimeters, greatest horizontal diameter, 95 millimeters; common thickness of lamella about 5 millimeters, extreme basal portion of sponge about 1o millimeters thick. Above, the lamella diminishes to a fairly thin margin. Both surfaces of the sponge are corrugated with more or less parallel ridges, about 2 millimeters apart, radial to the margin, converging and dying away toward the base. These are thickly beset with small conuli; between the ridges the surface is smooth. The extreme basal part, lacking well-defined ridges, is irregularly conulose. The surface is hispid, with spicules that project at the conuli. Color bright orange red. Consistency firm, but not rigid, fairly elastic, of the kind known as car- tilaginous. There is no difference between the two sides of the lamella. A few scattered inconspicuous oscula about 1 millimeter in diameter were found, and open pores, about 40 u in diameter, occur scattered over both surfaces. From the pores small pore canals pass through the outer stratum of the ectosome into subdermal canals, which extend parallel to the sponge surface. Characteristic subdermal canals measure too to 400 uw in diameter, the smaller sizes being the more common. From them canals lead more or less radially into the interior. Large canals are not abundant in any part of the sponge. The ectosome is 162 BULLETIN OF THE BUREAU OF FISHERIES. collenchymatous and about 250 to 350 u thick. Its superficial stratum is fibrous; that is, the cells are elongated and rather compactly arranged parallel to the surface. From the ectosome wide tracts of collenchymatous tissue pass into the interior, marking out the pathways of the main canals. Spicules.—The megascleres (Pl. LXVI, fig. 56a toc) are styles which fall in two very distinct classes, although intergrades can be found. (1) A stouter form (fig. 56a), smooth, cylindrical, usually slightly curved; the characteristic spicule of the peripheral skeleton. In the upper part of the sponge this spicule generally measures 400 to 600 u by roto 124. Inthe lower basal part of the body it is larger, reaching 7oou by 20. The oxeate modi- fication sometimes occurs (fig. 566). Intermediate forms, between style and oxea, with irregular, imperfect ends, also occur (fig. 56c, d, e). The stouter form of style is present in small number in the mesial skeleton, sometimes projecting from the individual fibers. A short form, often bent, about 200 to 350 u by 16 to 20 uw occurs with some frequency. (2) A slender form of style 400 to 500 u by 3 to 8 uw, the commonest thickness being about 4 to 6 u, is the characteristic spicule of the axial (mesial) skeleton. This spicule is smooth, cylindrical, usually slightly curved, sometimes exhibiting more than one bend. The oxeate modification occurs. Skeletal framework.—The framework includes an axial portion, which, since the sponge is essentially lamelliform, is better designated mesial, and a peripheral portion. The mesial skeleton is a reticulum made up of longitudinally coursing (i. e., extending from base to free margin of sponge) fibers, which branch and anastomose (Pl. LXV, figs. 46 and 47). In the basal part of the sponge this portion of the skeleton becomes quite thick; everywhere it occupies all but the superficial zone of the body. The fibers are cylindrical and plurispicular, with abundant spongin; the spicules arranged more or less longitudinally. Common range in thickness of individual fibers, 4o to 100 W. The peripheral skeleton consists chiefly of obliquely radial fibers, into which the mesial skeleton is produced (figs. 46 and 47). They are slender in the marginal region of the lamella, but become thicker below, in the basal region becoming strong and plumose. ‘The fibers include but little spongin, only enough to hold the spicules together. The radial fibers branch to some extent, and at the distal end the spicules fray out, forming a bunch or bunches, generally divergent in character. These outermost spicules are embedded in spongin only at their very base. The radial fibers extend into the conuli. If, as often happens, the conulus is subdivided into secondary conuli, each of these receives a spicule or two or a bunch of spicules. The peripheral skeleton includes also radial megascleres, which beset the mesial skeleton on its outer surface, between the radial fibers. They extend out toward the surface, and are especially con- spicuous in the depressions between the ridges or other conulose elevations. These spicules belong to the stouter form of style. Vosmaer has recently (1912) made an attempt to establish the genera Axinella, Phakellia, Acanthella, and Raspailia on more definite anatomical grounds. He finds that a new genus, Phacanthina, must be made for Schmidt's Acanthella obtusa. Vosmaer finds that the species which he has studied, representing the above five genera, differ in definite points as to the character of the skeletal framework, and he therefore concludes that, in the diagnosis of the genera, the character of the framework should play an important part. How well the numerous species hitherto grouped under the above genera, when studied anatomically, will fit into the five skeletal schemes described by Vosmaer is, of course, a question. But there is no doubt that the distin- guished spongologist, whose recent death entails such a loss upon zoological science, should be followed in the resolve to learn more definitely about the skeletal framework of these sponges and to use the data in classification. According to Vosmaer’s generic schemata, Acanthella and Phacanthina lack the radial fibers (extraaxial funiculi) which are found in Axinella, Phakellia, and Raspailia. The Beaufort sponge would therefore fall in Phakellia. But it seems not impossible that SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 163 some of the peculiarities of the skeletal framework which Vosmaer has brought out in his stimulating paper are such as distinguish species rather than genera. In listing the sponge under Acanthella we rely for the present on the assumption that the cartilaginous consistency, and especially the aculeate surface, when coupled with a lamellate form and the general type of Axinellid skeleton, are signs of kinship. Species of Acanthella, in the usual sense, are known from the Red Sea: A. aurantiaca Keller (Keller, 1889-91; Topsent, 19066; Row, 1911); Mediterranean: A. acuta Schmidt, A. obtusa Schmidt (O. Schmidt, 1862; Topsent, 1901a); Australian waters: A. stipitata Carter, A. fenuispiculata Dendy (Dendy, 1896); Gulf of Manaar: A. carteri Dendy, A. flabelliformis Keller (Dendy, 1905); Torres Straits: A. pulcherryma Ridley (Ridley, 1884; Ridley and Dendy, 1887); Aru and Kei Islands: A. euctimena Hentschel, A. sp. Hent- schel (Hentschel, 1912). It would seem that Acanthella flourishes in warm waters, although Vosmaer some time ago (1885) referred a sponge taken in the Arctic Ocean, A. multiformis, to this genus. KERATOSA. Family DARWINELLIDA (APLYSILLIDZ: auct.). Keratosa with eurypylous and large flagellated chambers; with a skeleton com- posed of separate horny fibers that ascend from the base of the sponge and are simple or branched, or the skeleton may be reticulate; spicules of spongin may also occur; the spongin fibers contain a medulla or pith and lack foreign mineral particles. Aplysilla F. E. Schulze. Incrusting or lamellar forms; skeleton nonreticulate, composed of separate fibers; without spongin spicules. Aplysilla longispina, n. sp. (Pl. LXIV, fig. 42; Pl. LXV, figs. 45, 48; Pl. LXVI, fig. 64a, 6, c, d.) One large, incrusting specimen taken on the piles of the Morehead railroad pier, near the town end, just below low-water mark. Sponge (Pl. LXIV, fig. 42) 1 to 20 millimeters thick, covering an area 150 millimeters in diameter, apparently with unlimited lateral growth. The surface is covered with numerous slender, sharp conuli several millimeters high (the range is 1 to 5 millimeters, the common height about 3 millimeters) and a variable distance apart (often about 3 to 5 millimeters), frequently but not always connected by thin, sharp-edged ridges or folds. Conuli in general simple, but not uncommonly bifid or trifid at the apex. Surface very uneven because of numerous ascending portions, all of which are low, irregular, and so vaguely delimited as not to merit the name of lobes. A few oscula, 1 to 3 millimeters in diameter, are scattered over the surface. Color, sulphur yellow, turning instantly in alcohol to an indigo blue. Sponge soft and elastic; interior cavernous (Pl. LXV, fig. 45). The skeletal fibers are simple, or branched somewhat in elk-horn fashion (Pl. LXVI, fig. 64a, b,c). At the under surface of sponge they expand into thin, basal, horny plates, which sometimes, at any rate, connect with one another. ‘The fibers extend vertically upward into the conuli, reaching the apex of the latter (Pl. LXV, fig. 45). The simple fibers support each a single conulus, and each terminal division of a branched fiber supports a conulus or one of the subdivisions of a primary conulus. The more complex fibers thus support a number of conuli. The fibers measure 100,to 250 w in diameter just above the basal plate and 15 to 30 u near the tip. The pith, which consists of successive thimble-shaped segments of varying length, forms about half of the fiber in the basal portion. The spongin wall is clearly stratified (Pl. LXVI, fig. 64d). The fibers are generally smooth and have no inclusions, but occasionally foreign bodies are found in them. 164 BULLETIN OF THE BUREAU OF FISHERIES. The surface of the sponge is formed by a dermal membrane 20 to 4o p thick, the body of which consists of not very closely packed mesenchyme cells, which are elongated parallel to the dermal surface and the neighboring spaces. The dermal membrane is traversed by closely set, short, radial canals, 75 to 125 win diameter and about 3o to 60 p apart. These open internally into large subdermal spaces. At the outer dermal end these canals are now closed in by thin membranes, each of which in life is probably perforated by several pores. In the actual specimen the pores are closed. The subdermal spaces are wide just beneath the dermal membrane, but, descending vertically from the surface, they become gradually narrower, terminating in the inhalent canals, which are not marked off by any definite limit from the subdermal spaces. The subdermal spaces and the inhalent canals open directly by small prosopyles, about 4 to 8 « in diameter, into the flagellated chambers, which are generally, but not invariably, longer than wide, 80 to 130 # long, 60 to 100 # wide, and which open directly by wide mouths into the exhalent canals. The Beaufort sponge is probably to be looked on as a migrant from the coast farther south. A Bahama species of Aplysilla is known, A. compressa (Carter), but this is an erect lamellar form (Lendenfeld, 1889, p. 704). The Mediterranean species, A. sulfurea F. E. Schulze, recorded also from Australian seas, the European coast of the North Atlantic, and from Juan Fernandez (Thiele, 1905, p. 488), resembles the Beaufort sponge in color, the yellow turning to violet in alcohol (Topsent, 1904, p. 56), or gradually becoming blue as the sponge dies in the air (F. E. Schulze, 1878). The conuli are much lower than in the Beaufort sponge, only 0.5 to 1 millimeter high and about 1 millimeter apart. The Beaufort sponge is nearer to an Australian form, A. violacea (Lendenfeld, 1889, p. 704). ‘This is an incrusting species with unlimited lateral growth and conuli about as high as in the Beaufort species. The conuli are, however, more closely set, and the natural color is dark violet. Moreover, there is a basal spongin plate which contains large sand grains, and the flagellated chambers are smaller (60 to 100 » long by 30 to 4o # wide) than in A. longispina. Recent writers (Topsent, 1905; Dendy, 1905 and 1916; Row, 1911; Hentschel, 1912) do not separate the Darwinellide and Aplysillide of Lendenfeld’s monograph (1889), but combine them in one family, for which Topsent uses the heading, Darwinellide, thus following in essentials Merejkowsky (1879) and Vosmaer (1887). The other authors cited above use the name Aplysillide. Topsent is obviously in the right, since the type genera of the families combined, Darwinel/a F. Miiller and Apiysilla F. E. Schulze, were established in 1865 and 1878, respectively. Family SPONGELID/. Keratosa with eurypylous and large flagellated chambers, with a skeleton com- posed of separate horny fibers that ascend from the base of the sponge and are simple or branched, or the skeleton is more commonly reticulate. Horny fibers without pith, generally containing abundant foreign mineral particles. Skeleton may be reduced, then consisting of foreign particles usually held together by a little spongin, but the latter may be absent. Pleraplysilla Topsent. Thin, incrusting forms, with low conuli, supported by spearate areniferous fibers ascending from base of sponge and characteristically undivided, SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 165 Pleraplysilla latens n. sp. (Pl. LXIV, fig. 4o.) Sponge occurs in the shape of thin, colorless incrustations on oyster shells, commonly along with Microciona prolifera. Under the piers along the ‘‘town front’’ proved to be the best collecting ground. ‘The sponge is not nearly so abundant as Microciona, but is not rare. It is, however, inconspicuous. It contains abundant embryos during July and August. Surface covered with small, sharp conuli, commonly o.5 to 1 millimeter apart. ‘Thickness of sponge from base to apex of conulus about 1 millimeter, usually something less; body of sponge itself about one-half that thickness. Upper ends of the skeletal fibers extending into conuli are conspicuous, reflect- ing the light. The fiber in some cases extends an appreciable distance beyond the substance of the conulus; but this may be an effect due to contraction. The dermal membrane in stained preparations exhibits narrow bands, which prove to be linear thickenings about 20 » wide, due to aggregations of mesenchyme cells on the under surface of the mem- brane. The mesenchyme cells are elongated in the direction of the bands. The bands radiate from the apices of the conuli, often about eight from a conulus, soon branching and passing into a network of similar bands which occupy the sides of, and the areas between, the conuli. The meshes of this net- work are polygonal, irregular in size, the diameter ranging from 70 4 to 250 u. The network is easily seen with a lens in a lightly stained preparation in alcohol. The bands are sometimes so arranged that primary meshes of the reticulum are subdivided. ‘The meshes themselves are riddled with small pores 12 to 24 4 in diameter. In the actual specimens examined the pores were open in some regions, closed in others. The dermal reticulum of bandlike thickenings just described is a structure similar to that found in Aplysina aerophoba, Aplysilla sulfurea, species of Spongelia, and other horny sponges (F. E. Schulze, 1878, 1879a). It is of the same general character as that found in A plysilla longispina nobis (P1. LXV, fig. 48), but in the latter species the bands are relatively very thick and strongly developed, so much so that they constitute lateral walls of distinct, though short, canals, which may be said to traverse the dermal membrane radially. Beneath the dermal membrane there is a nearly continuous subdermal space from which canals, presumably afferent, pass vertically downward into the sponge interior. These canals are numerous, often about 0.5 millimeter apart, although there is no regularity in their distribution. ‘The mouths of the canals, reaching-about 300 » in diameter, are easily seen through the dermal membrane. With the lens they look like surface apertures and give to the sponge a porous appearance. Numerous tangential canals, presumably efferent, extend just below the dermal membrane. ‘These are long and branching, the larger about 0.5 millimeter wide. The membrane covering them is without conuli, and lacks or nearly lacks the system of bandlike dermal thickenings. It is in general aporous, but a few scattered pores occur of about the same size as those found elsewhere. Open oscula 2co to 250 m in diameter were observed scattered over the surface of the sponge. The flagellated chambers are longer than wide, commonly ellipsoidal, often about roo to 120 by 7oto gow. They may be larger, reaching 175 by 100 uw, sometimes slightly curved in the direction of the greater axis. Still longer tubular chambers occur here and there, sometimes with indications of branching. ‘These bespeak the primitive nature of the canal system. ‘The chambers are perforated by numerous prosopyles 8 to 10 » wide, and open by wide apopyles directly into the efferent spaces. The chambers are abundant below the general subdermal space, except in the regions of the main tangential canals, below which there are, however, some. The skeleton consists of simple independent fibers, 40 to 60 u thick at about the middle, ascending from the basal surface of the sponge into the conuli. They are made up of spongin and mineral particles, the latter including sand grains, fragments of sponge spicules, occasional entire spicules, and diatom shells. The mineral particles, except at the base of the fiber, nearly or completely fill it. Close to the base the mineral matter tapers away to a thin core, leaving the surrounding spongin very evident. In this region and at other points also along the course of the fiber it is possible to see that the spongin is laminated. At the very base the fiber expands into a thin basal plate of spongin. The fibers some- times extend vertically, or nearly so, from base to conulus. But usually they extend obliquely from the base upward, often occupying throughout a large part of their course a more or less horizontal position. They are sometimes fairly straight, but frequently curved or bent. They are characteristically simple, neither branching nor connecting. Rarely a fiber is found with a lateral branch; and, again, rarely two fibers may come in contact and fuse, thus producing the appearance of a fiber that divides basally. 166 BULLETIN OF THE BUREAU OF FISHERIES. Topsent some years ago (1905, p. CLXXXV) described a new and interesting sponge for which he created the genus Pleraplysilla. Topsent’s sponge, P. minchini, was dredged off the French channel coast at a depth of 30 meters. The sponge is incrusting, about 1 millimeter thick, except in spots where the thickness reaches several millimeters. The color is chocolate. The largest specimen measured 25 centimeters in diameter. The surface is beset with conuli 1.2 to 2 millimeters or more apart. ‘The skeletal fibers are characteristically simple but, especially in the thicker parts of the sponge, may send out two or three branches; they are too to 110 p» thick below. The flagellated chambers are eurypylous and measure about 90by 35 uw. The Beaufort sponge is evidently another species of this genus. For Pleraplysilla and another sponge Igernella (Darwinella) joyeuxt, from the Gulf of Mexico, Topsent, Joc. cit., creates the family Pleraplysillide. Igernella having horny spicules would be a good Darwinella if the fibers of its skeletal reticulum were not areniferous. But this latter characteristic excludes it from the Darwinellide (Aplysillidee of some). Its horny spicules, on the other hand, exclude it from the Spongelide. The sponge is an intermediate between the Darwinellide and Spongelide, and Topsent’s family provides a place for it, although, if Pleraplysilla is removed, as we suggest (see below), the name of the family will have to be changed. Pleraplysilla, while it will not go in the Darwinellide because of its areniferous fibers, can not, it seems to us, be excluded from the Spongelide. It takes its place at the base of the latter family, its very simple fibers leading up to the more complicated, but still independent, ones of Spongelia spinifera. It is generally recognized that the separation between the Darwinellide and the Spongelidz is not a sharp one. Dendy (1905, pp. 203, 207) points out that Spongelia spinifera F. E. Schulze and Megalopastas Dendy are intermediate forms. Igernella and Pleraplysilla are also intermediates, although not intermediate in respect to the same points. Family SPONGIDA. Sponges with small flagellated chambers, 20 to 50 u wide, and a skeleton, generally in the shape of a reticulum, composed of solid or pithed, horny fibers. Subfamily STELOSPONGINA. Spongide in which main fibers and connectives are generally distinguishable in the skeletal reticulum. The main fibers may be simple but are generally more or less fascicular. Between the fascicular fibers, or between the simple main fibers in species without fascicles, the skeletal meshes are much larger than in the Euspongina. Hircinia Nardo. Stelosponginz with filaments in the ground substance and in which the connectives are characteristically attached to the main fibers by diverging roots which extend along the main fiber in one plane. Hircinia ectofibrosa, n. sp. (Pl. LXIV, figs. 41, 43, 44; Pl. LXVI, fig. 61.) Taken several times on Fort Macon beach after moderate gales; six specimens available for study; probably growing on the “Fishing Bank’’ off Beaufort Inlet and in similar places, and to be looked on as an outlying member of the Florida-West Indian fauna. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 167 The chief characteristics of the species, as brought out through the comparison of specimens, no two of which agree in detail, are: The predominantly simple character of the main fibers and the con- nectives; the existence of dermal connectives; the tendency for the outer ends of the radial fibers to become fascicular through the extension of, and union between, the roots of the surrounding dermal connectives. The shape of the body in most of the specimens is platelike, the plate probably standing more or less vertically; the platelike body may be produced into lobes. The platelike shape is not universal, for in the type specimen (Pl. LXIV, fig. 41) there is simply a common basal part dividing above into a few subcylindrical or flattened lobes. The specimens vary in total height from 50 to 130 millimeters; in thickness from 10 to 35 millimeters. The sponge is beset with sharp conuli connected by thin interconular ridges, which divide the surface into rounded or polygonal, depressed areas. The conuli project 1 millimeter or less above the ridges but 1.5 to 3 millimeters above the level of the depressed areas. They are 2 to 4 millimeters apart, these distances representing usually the diameters of the interconular depressed areas, but the bounding ridges are sometimes absent, with the result that interconular areas are produced larger than the normal. The ridges themselves vary in height from a fraction of 1 millimeter to 2 millimeters. The specimens studied all being beach specimens, although fresh and in good condition, it is quite probable that in them the conuli and interconular ridges are sharper, and the depressed areas deeper, than in the living sponge, owing to contraction incident upon partial drying. Over a part of one specimen the radial fibers (see below) project from the summit of the conuli. But this is a condition obviously pathological (in a wide sense) and of no classificatory value. The conuli are generally vertical to the surface, but on parts of several of the platelike specimens they incline obliquely upward toward the free margin of the sponge. ‘Thus several interconular ridges, together with the depressions between them, are combined along lines that radiate toward the free margin. ‘This leads to the appearance of ridges and furrows, radial to the margin, that may be 15 milli- meters long, a type of surface architecture which approaches that shown in some of Hyatt’s specimens of H. campana (Hyatt, 1877, Pl. XVII, fig. 28). In such regions the dermal connectives (see below) in a ridge are combined to form continuous fibers which extend the length of the ridge; these are con- nected by transverse, simple fibers, and thus a ladderlike dermal skeleton is produced. Foreign mineral particles are present in considerable abundance on the surface and in the ectosome, which thus sometimes to a depth of 1804 appears dense and whitish as compared with the light yellowish- brown interior. The particles include the usual sand grains, spicule fragments, and foraminifer shells or bits of the same. They are scattered; that is, are not abundant enough to be massed together so as to form a continuous sand cortex. The particles for the most part are small, but in several specimens large sand grains, up to 250 u in diameter, are abundant among the smaller bits. A surface reticulum consisting of pore areas separated by trabecule containing closely packed sand grains is not present in these specimens. But this reticular appearance, which is so common in horny sponges and, because of the sand grains in the trabeculz, so conspicuous in some Hircinias, is, as has been shown (Wilson, 1902, p. 405), greatly influenced by the physiological state; that is, whether the pores are open or not. There are in several specimens of the present species indications that in the active state pore areas will be found, 150 to 250 u in diameter, each containing one or more pores 25 to 70 » in diameter, and separated by interareal trabecule of thicker dermal membrane full of mineral particles. Actually only a few pores are open. Subdermal cavities 2 to 3 millimeters wide are abundant, more so in some specimens than in others. The sponge interior is porous, with canals upto about 3 millimeters in diameter. The flagellated cham- bers measure about 36 by 28 4. Oscula, 1 to 4 millimeters in diameter, are scantily scattered over both surfaces in the tabular specimens. The filaments (Pl. LX VI, fig. 61) are very abundant. They are about 6 » in diameter at the middle, 3 to 4m near the ends. The terminal enlargements are oval and about 9 » thick. The filaments are smooth and for the most part colorless and unspotted, but in one of the specimens some of the filaments are “‘spotted,’’ others not. The ‘‘spotted’’ or ‘‘unspotted’’ condition is obviously of no classificatory value. ‘The filaments in some regions, especially around some of the canal walls, are arranged in more or less distinct tracts. The surface color varies from whitish to a dull purple. The natural color has probably In part faded out. The sponge is firm, but compressible and elastic. 168 BULLETIN OF THE BUREAU OF FISHERIES. The skeleton includes main fibers and connectives. The main fibers ascend in the middle of the plate, or lobe, and branch, the branches curving outward in the usual way as radial fibers, which ter- minate in the conuli. The ascending and radial fibers are alike, and both are referred to here, in accord- ance with the general usage, as main fibers. The radial fibers, in macerated skeletons, are 1.2 to 2 millimeters apart. The main fibers (Pl. LXIV, figs. 43, 44) are characteristically simple, solid fibers, 100 to 200 y thick, well filled with mineral particles (sand grains, bits of sponge spicules, and foraminifer shells), all comparatively small in size, there being no large sand grains or shells to swell out and distort the fiber. The spongin is stratified. In some specimens, but not in others, the outermost ends of the main fibers, and the dermal connectives as well, are composed of a much paler spongin than the rest of the skeleton, in which the spongin is yellow. This difference is probably associated with some individual difference in growth activity. Whenever the main fibers appear in any degree fascicular, this is due to one of two causes, as follows: (1) Two or even three main fibers may be closely approximated, or a main fiber may branch obliquely, the two or three branches continuing to run more or less parallel and close together. Between such fibers or such branches, respectively, the connectives are, of course, very short, and the several fibers, together with their connectives, constitute a compound fiber such as is characteristic of Stelo- spongia. Such compound fibers have a total thickness of about 0.5 millimeter. Thiscondition is found here and there in the specimens studied, but is comparatively rare. In the literature on the Stelo- spongine when the term “fascicular fiber’’ is used, writers seem usually to have in mind a compound fiber of this kind. (2) A connective unites with a main fiber by several roots, the middle about transverse to the fiber, the upper and lower oblique to the fiber. Thus the roots of a connective attach themselves to a considerable extent of the main fiber, forming altogether a sort of triangular plate. If now, as often happens, several connectives attach to the same immediate region of a main fiber, but on different sides, the main fiber in that region is surrounded by several sets of roots, and thus may appear “‘fascic- ular,’’ although in a different sense from that understood under (1). Between any two successive sets of roots which meet it the main fiber is, as a rule, obviously simple (fig. 44), but in places the roots spread up and down the fiber so far that successive sets meet one another. The main fiber then appears not as an independent fiber, but asan axial tract, distinguishable because of its mineral contents, extend- ing through an elongated, and close, reticulum. Such a condition is found here and there in the inte- tior of the sponge. It is commoner in the outer layer of the body, between the dermal connectives and the most superficial of the inner connectives, which usually lie about 1 millimeter below the former. In its outermost part, within a conulus, the main fiber often remains simple; but in this region the fibers vary considerably, and the variants need to be described. ‘They are as follows: (2) The outer end, intraconular portion, of a main fiber is, as said, often simple and so full of mineral particles as to show very little spongin. In a typical case it extends 500 » beyond the attaching roots of the dermal connectives and ends in a slight enlargement, each dermal connective meeting the fiber by several roots. The terminal portion of the fiber may show a perforation or two, due probably to the fact that mineral particles here have only recently been surrounded by the spongin of the fiber. (2) The roots of the dermal connectives may extend very obliquely upward (and downward) along the main fiber. These roots, thus entering a conulus from several sides and extending up toward the apex, become interconnected by short fibers, and so may give rise to a very perfect trellis, suspended like a tent, as it were, from the uppermost part of the main fiber. This is, perhaps, the commonest condition of the intraconular portions of the main fibers in the specimens studied. (3) In exceptional cases the main fiber, near its outer end, may divide, the two branches extending into the same conulus. They are connected in a close and complex way, the connectives themselves being so united as to constitute reticula. Terminally the two branches may end in a common enlarge- ment, which maceration shows is a spongin reticulum very full of mineral fragments. In such a case the intraconular skeleton is fairly to be classed as ‘‘fascicular’’ in the sense of being a compound, Stelo- spongia-like fiber. A variant of this condition is found where two or three main fibers which, in the periphery of the sponge, at least, are not branches of a common fiber, converge and enter the same conulus, within which they are united by short connectives. The most superficial, or dermal, connectives lie in the ectosome just below the dermal surface, passing from the outer end of one main fiber to that of another. Except where interconular ridges are not developed, they lie in and close to the free edge of the latter. In general they are single, simple SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 169 fibers connecting with the main fibers by several roots. They are commonly united by other dermal fibers that may be called interconnectives, but these are few in number and irregular in distribution. The result is that the dermal skeletal reticulum has, in general, large meshes that approach the squarish or rectarigular shape. Characteristic meshes measure 1.6 millimeters by 1.9 millimeters, 2.2 millimeters by 2.5 millimeters, 2.5 millimeters by 2.5 millimeters, 2 millimeters by 4 millimeters. The dermal connectives in some specimens are very different in appearance from the internal connectives, in that they are slenderer than the average internal connective and of very pale spongin well filled with min- eral particles. In other specimens they differ from the internal connectives only in containing more mineral particles, while in still other specimens the dermal connectives do not differ in appearance from the internal ones. The difference in appearance between dermal and internal connectives is thus inconstant. It is, perhaps, in part correlated with a difference in growth activity. In one specimen the dermal connectives depart, over much of the surface, from the type. Typical connectives are developed, but the surface as a whole offers the following deviations from the type: (a) Instead of a few interconnectives, a comparative abundance of fibers develop between and around the principal connectives, thus producing reticula which encroach upon the interconular areas. Such reticula may be very fine. (b) In this specimen as in some others a good many large sand grains are scattered through the ectosome and on the surface; but, whereas in the other specimens they have not been incorporated in the skeletal fibers, in this specimen a great many of the dermal connectives include large grains that measure up to 250 4 in diameter. The grains lie in a longitudinal series in the connective and are surrounded by a thin layer of spongin. They may form a continuous series or be separated by inter- vals in which the connective remains of the usual thickness, about 4oto 50u. Such fibers resemble those of the species grouped together by Lendenfeld under the subgenus Psammocinia (1889, p. 579)- (c) A further complication is present in that some of the dermal connectives that have incorporated large sand grains are fascicular. Such ‘‘fascicular”’ fibers consist of several simple connectives that lie close together and are interconnected. Sometimes all the longitudinal components of such a fiber apply themselves to the same sand grain. Fascicular connectives of this kind may be 350 u wide, the constituent being strands only 20 p thick. The peculiarities of the dermal skeleton in this specimen are probably no more than individual differences. The specimens form a series, at one end of which are those without large sand grains; in the middle, those with large sand grains in the ectosome but not in the skeletal fibers; at the opposite end, the specimen in which many dermal connectives have incorporated the sand grains in question. The internal connectives are composed of yellow spongin and in general are without, or have only scanty, mineral contents. Close to the surface they may exceptionally contain more, but even then the mineral contents can not be said to be abundant. They range in thickness from about 35 w to 175 yu. The connectives are characteristically simple, the meshwork correspondingly coarse. The skeletal meshes in the macerated skeleton, which with the most careful treatment shrinks in some degree, may reach 2 millimeters in diameter. Common sizes for the largest meshes are: Width 1.5 millimeters, with aradial diameter of 1.2 millimeters; width 1.2 millimeters, with a radial diameter of 1 millimeter; width 1.2 millimeters, with a radial diameter of 500 to 600 uw. In the case of such meshes neighboring connectives are not united together. But very commonly neighboring connectives are united together by a few other fibers—‘“interconnectives,’’ as they may be called. The size of the mesh is thus cor- tespondingly reduced, although characteristically it still remains large, typical diameters ranging from 300 to 600 yu. The interconnectives often become so numerous and complex that they, together with the connec- tives, form reticula that extend between the main fibers, thus making an approach to the condition characteristic of Hircinia fetida (Schulze, 18796, Pl. III, fig. 3), although the reticula in question are coarser and more irregular than in the latter species. Position of H. ectofibrosa in the genus.—The Beaufort species is in that group of forms which center around the Mediterranean sponges described by F. E. Schulze as Hircinia variabilis, in which the main fibers are simple or only slightly fascicular and the connectives characteristically simple. Lendenfeld combines these forms into a subgenus, Euricinia. In his definition of this subgenus the following clause must now 170 BULLETIN OF THE BUREAU OF FISHERIES. be omitted: “but no large sand grains joined by slender short fibers occur” (1889, P- 554)- In the possession of a dermal skeleton the Beaufort species appears to be nearly unique in the subfamily. Lendenfeld says (loc. cit., p. 477) that the species of Stelo- spongia are destitute of a special dermal skeleton, by which he must mean one that lies in the ectosome, close to the surface and above the level of the subdermal cavities. So, too, the described species of Hircinia, in general, lack a dermal skeletal reticulum, which is, however, present in H. (Oligoceras) conulosa (Ridley) (Ridley, 1884, p. 599; Lendenfeld, loc. cit., p. 535). The presence of large sand grains in the skeletal fibers is the central fact on which Lendenfeld bases his subgenus Psammocinia (loc. cit., p. 579). As to whether Psam- mocinia is a natural group or an assemblage of phenotypes we are not in a position to form an opinion, although it is certain that the mineral content of the skeleton is exceedingly variable, both in total amount and kind, in what must be regarded as a single species. F. E. Schulze long ago pointed out how variable is the amount of mineral content in the ectosome of H. variabilis (1879b, p. 14). The resemblance of H. ectofibrosa to some of the Mediterranean specimens of H. variabilis involves surface details. "These are the sponges now assigned to var. hirsuta, in which the conuli are high and sharp and often in rows: “‘gew6hnlich in kurzen unre- gelmassigen Kammen, welche Bogen bilden und in einander tibergehen” (Schmidt, 1862, p. 33). They are evidently very similar to the Beaufort species. The Beaufort sponge is doubtless an outlying member of the Florida-West Indian fauna. Several species of the genus have already been recorded from the Florida- West Indian waters: Hircinia campana (Lamarck), H. arbuscula (Schmidt), H. acuta (Duchassaing et Michelotti), H. cartilaginea (Esper), H. purpurea Hyatt, by Hyatt (1877); several under “‘ Polytherses’’ by Duchassaing and Michelotti (1864, p. 67); H. caracensis Carter (Carter, 1882, p. 273), H. twbulosa Carter (Carter, 1884, p. 203); H. purpurea Whitfield, and H. atra Whitfield (Whitfield, 1901); H. acuta (Duchassaing et Michelotti), H. variabilis F. E. Schmidt, H. jetida (Schmidt) var. cuspidata Wilson, by Wilson (1902). Not all of these species are recognizable. One of the West Indian forms just recorded offers resemblances to the Beaufort species. This is H. campana var. fixa Hyatt (Hyatt, 1877, p. 546, Pl. XVII, fig. 28). The case of H. (Spongia) campana (Lamarck) as occurring in the West Indies is as follows: Duchassaing and Michelotti (1864, p. 68) identified certain forms as belonging to this species. Hyatt (loc. cit.) grouped under the same heading a variety of West Indian sponges. He tells us that the variation in his specimens is great, involving not only habitus and size of the interconular depressed areas, but the skeleton also. In the absence of more detailed structural data it is uncertain how far Hyatt was jus- tified in grouping these forms together. With respect to some of them, varieties tyfica, fixa, and felix, his account intimates that the main fibers are fascicular, in the sense of being compound fibers. Lendenfeld (loc. cit., p. 561) refers others of Hyatt’s speci- mens, var. columnaris to H. variabilis, evidently concluding (on what grounds is uncer- tain) that the main fibers in these are simple. With respect to H. campana, we venture to say that possibly the vase shape is assumed by several West Indian Hircinias, or there may be a West Indian species (H. campana) which under certain conditions becomes vase shaped, but which may be of SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 171 almost any habitus; in which the skeleton is very variable and imperfectly known; and in which the interconular areas vary greatly in size, from 2 to 10 millimeters in diameter; the species remaining recognizable in spite of its variability. Cotes and Yarrow (1878) record under this name a specimen collected in the Beaufort region. Schriidt (1870, p. 30) had also before him West Indian specimens which he iden- tified as H. campana, but he frankly confesses that the wealth of Hircinia “forms” in the West Indian waters is so great that he can not divide them into species. This entanglement of “‘forms” still waits for its satisfactory analysis. Familiarity with considerable numbers of the living sponges in several localities and some breeding experiments are doubtless necessary for real success. Lendenfeld (loc. cit., p. 569) classes Spongia campana as a Sarcotragus (= Hircinias with distinctly fascicular main fibers), and to this species assigns the vase shaped and flabellate sponges from the West Indies and Florida named Polytherses campana by Duchassaing and Michelotti (loc. cit.), Hircinia campana by Schmidt (loc. cit.), H. campana varieties typica and fixa by Hyatt (loc. cit.). Discussion of the genus.—Lendenfeld in defining Hircinia (loc. cit., p. 545) lays stress, justly, we think, not only on the presence of filaments but on the way in which the connectives attach to the main fibers. He says: “‘The fascicular nature of the connecting fibers and their mode of attachment to the main fibers by numerous diverg- ing roots, which extend in one plane, distinguish most species of Hircinia sufficiently from Stelospongia and from all other genera.” He goes on to say that there are forms without this peculiarity, but these are plainly very close to the sponges with the pecu- liarity. Possibly this reasoning justifies Lendenfeld’s inclusion under Hircinia of forms with dendritic fibers instead of a reticulate skeleton, such as Cacospongia collectrix Polejeff and Oligoceras conulosum Ridley. The fascicular main fibers of Hircinia deserve a few words. They are structures that are somewhat vaguely treated in the literature and probably are not always under- stood in the same sense. The fact that the connectives are united to simple main fibers by diverging roots, which may extend in one plane along the fiber for a consider- able distance, leads to the formation of a type of fascicular fiber different from that which is characteristic of Stelospongia (ante under description of this species). An examination of the literature indicates that it is widespread among the species now recorded under Hircinia. This type of fascicular fiber owes its existence, we believe, (1) to the extension of roots along the main fiber such that roots of successive connectives join one another, and (2) to the fact that several sets of roots which surround the main fiber at about one level but on different sides combine to form a close-meshed reticulum through the axis of which runs the original main (simple) fiber. ‘The fascicular fibers appear to have this character in H. favosa and H. fetida (Lendenfeld, loc. cit., pp. 571, 577), although Lendenfeld regards the fascicular fibers of this subgenus (Sarcotragus) as composed of several individual fibers joined at frequent intervals to one another (loc. cit., p. 533); that is, as compound fibers. But in H. javosa the sand grains are restricted to the most axial fiber of the fascicle, the whole structure appearing to consist of an axial fiber “surrounded by garlands of slender fibers” (loc. cit.). The structure appears to be the same in H. jatida. Again, in describing the connectives of several species of subgenus Sarcotragus, Lendenfeld (p. 534) says the roots of the connectives “appear as continuations of the fibers which form the garlands in the main fascicles.” 172 BULLETIN OF THE BUREAU OF FISHERIES. The size of the flagellated chambers has been adduced as a differential feature marking off Hircinia and Stelospongia (Lendenfeld, loc. cit., p. 484). The chambers are said to be larger in Stelospongia, the diameter being given as 41 to 48 w. But in Hircinia variabilis F. E. Schulze, the diameter may be as much as 4o p. In such a case this generic differential can not be applied, although it may hold for the majority of the species of the two genera. The fundamental character of the fascicular fibers appears to be constant in Stelo- spongia. They are compound fibers (Lendenfeld, loc. cit., p. 478; Pl. XXXI, fig. 7; Pl. XXXII, figs. 7, 8, 9, 10). Lendenfeld sometimes calls them “plexus bands’’ (PI. XXXI, figs. 4, 10, 12, 14; Pl. XXXII, figs. 9, 10). The constituent parallel fibers may be well apart, or may come together so closely as to fuse (Pl. XXXI, fig. 14). The width of the compound fiber thus varies greatly from less than 200 yp to several milli- meters. It is unimportant whether the several main fibers of a ‘‘fascicle’” arise as branches of a common fiber or not. Farther in the interior doubtless they often unite, although separate peripherally. Compound fibers of this kind, as we have seen, un- doubtedly occur here and there in Hircinia ectofibrosa, but in this species certainly, and probably in the genus at large, when the main fibers become fascicular the character- istic formative method practiced is that of incorporating the roots of the connectives. The characteristic fascicular fibers of the two genera are thus probably quite different structures. Since in Lendenfeld’s system both under Hircinia and Stelospongia there are forms with single main fibers, H. variabilis and S. (Cacospongia) vesiculifera, for example, we are driven back in the separation of these forms, and hence in the separation of the genera, to the presence or absence of filaments and of ‘‘root plates” formed by the divergent roots of the connectives. BIBLIOGRAPHY. BoWERBANK, J. S. 1864. Monograph of the British Spongiade. Vol. I. Ray Society. London. 1866. Idem. Vol. II. Ibid. 1882. Idem. Vol. IV. Ibid. 1872. Contributions to a general history of the Spongiadz. PartI. Proceedings, Zoological Society of London. London. Carter, H. J. 1882. Some sponges from the West Indies and Acapulco in the Liverpool Free Museum described, with general and classificatory remarks. Annals and Magazine of Natural History (5), IX. London. 1884. Catalogue of marine sponges, collected by Mr. Jos. Willcox, on the west coast of Florida. Pro- ceedings, Academy of Natural Sciences of Philadelphia, vol. 36, 3d series. Philadelphia. CovEs, Ex.iotr, and Yarrow, H. C. 1878. Notes on the natural history of Fort Macon, N. C., and vicinity (No. 5). Proceedings, Acad- emy of Natural Sciences of Philadelphia, vol. 30, 3d series. Philadelphia. Denny, A. 1894. Catalogue of noncalcareous sponges collected by J. Bracebridge Wilson, Esq., M. A., in the neighborhood of Port Phillip Heads, part 1. Proceedings, Royal Society of Victoria, vol. 7. Melbourne. 1895. Part 2, idem, vol. 8. Ibid. 1896. Part 3, idem, vol.9. Ibid. 1905. Report on the sponges collected by Prof. Herdman, at Ceylon, in 1902. In: Herdman Rep. Pearl Oyster Fisheries, Part III. London. 1916. Report on the noncalcareous sponges collected by Mr. Jas. Hornell at Okhamandal, etc. Williams and Norgate, London. DEsor, E. 1848. Two new species of sponges. Proceedings, Boston Society of Natural History, Vol. III. Cambridge, Mass. DucHAssAInG et MICHELOTTI. 1864. Spongiaires de la mer Caraibe. Haarlem. _ HENTSCHEL, E. 1g11. Die Fauna Siidwest-Australiens. Tetraxonida, 2 Teil. Fischer, Jena. 1912. Kiesel-und Hornschwamme der Aru-und Kei-Inseln. Abhandlungen der Senckenbergischen naturforschenden Gesellschaft, Bd. XXXIV. Frankfurt. 1g14a. Die Spiculationsmerkmale der monaxonen Kieselschwamme. Mitteilungen aus der natur- historischen Museum zu Hamburg, XXXI. Hamburg. 1914b. Monaxone Kieselschwimme und Hornschwamme der deutschen Siidpolar-Expedition IgoI-1903. Reimer, Berlin. Hyarrt, A. 1877. Revision of North American Porifere. Part2. Memoirs, Boston Society of Natural History, vol. 2. Boston. KELLER, C. ‘ 1889-1891. Die Spongienfauna des Rothen Meeres. Zeitschrift fiir wissenschaftliche Zoologie, Bd. XLVIII und LII. Leipzig. Kirxpatrick, R. 1goo. On the sponges of Christmas Island. Proceedings, Zoological Society of London. London. tge8. National antarctic expedition. Natural History, Vol. IV; Porifera Tetraxonida. British Museum. London. 110307°-—21——12 173 174 BULLETIN OF THE BUREAU OF FISHERIES. LAMBE, LAWRENCE M. : 1893. Sponges from the Pacific coast of Canada. Proceedings end Transactions, Royal Society of Canada, Vol. XI. Montreal. 1896. Sponges from the Atlantic coast of Canada. Ibid, Vol. II. Sect. IV, 2d series, 1896-97. Lerpy, Jos. 1889. The boring-sponge, Cliona. Proceedings, Academy of Natural Sciences of Philadelphia, vol. 41, 3d series. Philadelphia. LENDENFELD, R. von. 1888. Descriptive catalogue of the sponges in the Australian Museum, Sydney. Publication of the Australian Museum. 1889. A monograph of the horny sponges. Triibner & Co., London. Published for the Royal Society. 1903. Das Tierreich, 19 Lieferung. Tetraxonia. Friedlander und Sohn, Berlin. 1906. Wissenschaftliche Ergebnisse der deutschen Tiefsee-Expedition, Bd. XI. Die Tetraxonia. Fischer, Jena. Linpcren, N. G. 1898. Beitrag zur Kenntnis der Spongienfauna des malayischen Archipels und der chinesischen Meere. Zoologische Jahrbiicher, 2. Abtheilung fiir Systematik, Bd. XI. Jena. LunbsBeEcx, W. 1902. The Danish Ingolf-Expedition. Vol. VI,No.I. Porifera(part1.) Published by the Zoolog- ical Museum of the University, Copenhagen. 1gos. Idem. Porifera (part 2). MARSHALL, WILLIAM. 1880. Untersuchungen iiber Dysideiden und Phoriospongien. Zeitschrift fiir wissenschaftliche Zoologie, Bd. XXXV. Leipzig. RInD.ey, S. O. 1884. Spongiida. In: Report on the zoological collections made in the Indo-Pacific Ocean during the voyage of H. M. S. Alert, 1881-82. British Museum. London. Rip.eéy, S. O., and Denpy, A. 1887. Report on the Monaxonida collected by H. M. S. Challenger. In: Reports on the scientific results of the voyage of H. M. S. Challenger during the years 1873-1876. Zoology, Vol. XX. Edinburgh. Row, R. W. H. 1911. Reports on the marine biology of the Sudanese Red Sea, xix. Report on the sponges col- lected by Mr. Cyril Crossland in 1904-5. Part2. Noncalearea. Journal Iinnean Society. Zoology, Vol. XXXI. London. ScumipT, O. 1862. Die Spongien des adriatischen Meeres. Engelmann, Leipzig. 1868. Die Spongien der Kiiste von Algier. Engelmann, Leipzig. 1870. Grundziige einer Spongienfauna des atlantischen Gebietes. Engelmann, Leipzig. ScHu1ze, F. E. 1878. Untersuchungen iiber den Bau und die Entwicklung der Spongien. IV. Die Familie der Aplysinide. Zeitschrift fiir wissenschaftliche Zoologie, Bd. XXX. Leipzig. 1879a. Idem. VI. Die Gattung Spongelia. Ibid., Bd. XXXII. n. g. 18796. Idem. VIII. Die Gattung Hircinia Nardo und Oligocerasn.g. Ibid., Bd. XXXIII. 1904. Wissenschaftliche Ergebnisse der deutschen Tiefsee-Expedition, Bd. IV. Hexactinellida. Fischer, Jena. Sotuas, W. J. 1888. Report on the Tetractinellide collected by H. M. S. Challenger. In: Report on the scien- tific results of the voyage of H. M.S. Challenger during the years 1873-1876. Zoology, Vol. XXV. Edinburgh. THIELE, J. 1903. Kieselschwamme von Ternate. II. Abhandlungen der Senckenbergischen naturfors- chenden Gesellschaft, Bd. XXV. Frankfurt. 1905. Die Kiesel und Hornschwiamme der Sammlung Plate. Zoologische Jahrbiicher, Supple- ment 6, Bd. III. Jena. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 175 TopsEnT, E. 1889. Cliona celata ou Cliona sulphurea? Bulletin de la société zoologique de France, T. XIV. Paris. 1892. Contribution a 1’étude des Spongiaires de 1’Atlantique nord. Résultats des campagnes scientifiques accomplies sur son yacht par Albert 1°T, Prince Souverain de Monaco. Fasc. II. Monaco. 1894a Application de la Taxonomie actuelle a une collection des Spongiaires du Bane de Campéche et de la Guadeloupe décrite précédemment. Mémoires de la société zoologique de France, T. VII. Paris. 1894b. Une réforme dans la classification des Halichondrina. Ibid. 1896. Matériaux pour servir a 1’étude de la faune des Spongiaires de France. Ibid., T. IX. 1898. Introduction a 1’étude monographique des Monaxonides de de France. Classification des Hadromerina. Archives de Zoologie expérimentale et générale (3), VI. Paris. 1899. Documents pour la faune des Spongiaires des cétes de Belgique. Archives de Biologie, T. XVI... Paris-Bruxelles. Ig00. Etude monographique des Spongiaires de France. III. Monaxonida (Hadromerina). Archives de zoologie expérimentale et générale, T. VIII. Paris. tgora. Considérations sur la faune des Spongiaires des cétes d’ Algérie. Eponges delaCalle. Ibid., T. IX. rgorb. Résultats du voyage du S. Y. “Belgica’’ in 1897-98-99. Rapports scientifiques. In: Expedition antartique belge. Zoologie. Spongiaires. Buschmann, Anvers. 1904. Spongiaires des Acores. Résultats des campagnes scientifiques accomplies sur son yacht par Albert 1°, Prince Souverain de Monaco. Fasc. XXV. Monaco. 1905. Etude sur les Dendroceratida. Archives de Zoologie expérimentale et générale (4), T. III. Notes et Revue. Paris. 1g06a. Les Clavulides purpurines. Bulletin du Muséum d’histoire naturelle, T. XII. Paris. 1906). Eponges recueillies par M. Ch. Gravier dans la Mer Rouge. Ibid. 1909. La Coupe de Neptune, Cliona patera. Archives de Zoologie expérimentale et générale (4), T. IX; Notes et Revue. Paris. 1913. Spongiaires de 1’éxpedition antartique national écossaise. Transactions, Royal Society of Edinburgh, Vol. XLIX for 1912-13, Part III. Edinburgh. VeRRILL, A. E. 1874. Report upon the invertebrate animals of Vineyard Sound and the adjacent waters, with an account of the physical characters of the region. Supplementary paper XVIII to Com- missioner’s Report, U. S. Commission of Fish and Fisheries, 1871-72, Washington. Vosmaer, G. C. J. 1885. The sponges of the ‘‘William Barents’’ expedition, 1880 and 1881. Bijdragen tot de dierkunde, Aflevering 12. Amsterdam. 1887. Bronn: Die Klassen and Ordnungen des Thier-Reichs, Bd. 2. Spongien. Leipzig and Heidelberg. 1902. On the shape of some siliceous spicules of sponges. Proceedings, Sections of Sciences, Konin- klijke Akademie van Wetenschappen te Amsterdam. Amsterdam. 1908. Poterion a boring sponge. Ibid. 1909. On the spinispirz of Spirastella bistellata (O. S.) Ldfd. Ibid. 1911. Siboga-Expeditie. The Porifera of the Siboga-Expedition, II. The genus Spirastrella. Brill, Leiden. 1912. On the distinction between the genera Axinella, Phakellia, Acanthella a.6. Zoologische Jahrbiicher, Supplement XV, Bd. I. Jena. WHITFIELD, R. P. tgor. Notice of a new sponge from Bermuda and of some other forms from the Bahamas. Bulletin, American Museum of Natural History, vol. 14. New York. 176 BULLETIN OF THE BUREAU OF FISHERIES. Wison, H. V. 1902. The sponges collected in Porto Rico in 1899. Bulletin, U. S. Commission of Fish and Fish- eries, Vol. XX for 1900, 2d part. Washington. 1904. Reports on an exploration off the west coasts of Mexico, Centra} and South America, and off the Galapagos Islands * * * during 1891 * * * XXX. The sponges. Memoirs, Museum of Comparative Zoology, Vol. XXX. Cambridge, Mass. toro. A study of some epithelioid membranes in monaxonid sponges. The Journal of Experi- mental Zoology, Vol. IX. Philadelphia. 1911. Development of sponges from dissociated tissue cells. Bulletin, U. S. Bureau of Fisheries, Vol. XXX for 1910. Washington. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. EXPLANATION OF PLATES. [Pls. LVI-LVX, from photographs; Pl. LXVI, from drawings.] PiatE LVI. 1. Poterion atlantica. Vertical section, including cortex of outer surface and adjacent choano- some. X 21. 2. Clionacelata. Section through oyster shell showing included sponge trabecule extending up beyond the surface in shape of oscular papilla, which is cut tangentially. X 15. 3. Spirastrella andrewsii. Section vertical to the cloacal surface between oscula. % ts. 4. Cliona celata. Pore papilla. Expanded cap is shown cut in vertical section. The section, below, is tangential and shows the surface of the wall of papilla. > 4o. 5. Clionacelata. Onoyster shells. 1. 6. Spirastrella andrewsii. Side view of piece shown in fig. 7b. The dermal surface is at top of figure. Large incurrent canals, between which lie comparatively thin septa of sponge tissue, extend radially inward from dermal surface. 1. 7a. Spirastrella andrewsii. Small part of cloacal surface. Oscula contracted. 1. 7b. Spirastrella andrewsii. Small part of outer surface showing incurrent apertures. » 1, Pirate LVII. 8. Suberites undulatus. Whole sponge, from the side. Base of sponge, to the left in the figure. Dan 9. Suberites undulatus. From a section showing fibrous skeleton of the interior. ro. to. Suberites undulatus. Entire transverse section through a lobe. The dark patches in the choanosome represent the longitudinal skeletal fibers cut across. Io. 11. Suberites undulatus. From a transverse section through a lobe. Surface of sponge, above and to the left. X 15. 12. Reniera tubifera. Dermal surface of an oscular tube. 2r. Pirate LVIII. 13. Stylotella heliophila. Whole sponge. X 1. 14. Tetilla laminaris. Side view of a small specimen. X %. 15. Reniera tubifera. Whole sponge. 34. PuateE LIX. 16. Reniera tubifera. Longitudinal section through a branch, including both surfaces. 21. 17. Tetilla laminaris. From a section vertical to one of the flat surfaces of the sponge. The long fiber in the interior ascends toward the upper edge of the sponge. 1s. 18. Stylotella heliophila. Dermal membrane, in surface view. 15. 19. Stylotella heliophila. Section vertical to surface. 2r. PLATE LX. 20. Esperiopsis obliqua. Biseriate habitus. 14. 21. Esperiopsis obliqua. Chaliniform habitus. Part of a macerated specimen, including the base. X %. 22. Esperiopsis obliqua. From a longitudinal section through a branch of the macerated specimen shown in Fig. 21. X 21. 23. Esperiopsis obliqua. From a transverse section through a branch of the macerated specimen shown in Fig. 21. X 21. 177 178 BULLETIN OF THE BUREAU OF FISHERIES. Pirate LXI. Fig. 24. Phoriospongia osburnensis. Sponge incrusting on alcyonarian. % 1. Fig. 25. Phoriospongia osburnensis. From a section through the incrustation, vertical to the surface. Alcyonarian colony cut twice. X 28. Fig. 26. Lissodendoryx carolinensis. Whole sponge. X 34. Fig. 27. Lissodendorynx carolinensis. Papilla. 1s. Fig. 28. Lissodendoryx carolinensis. Interior of sponge. From a section vertical to the surface. 15. Pirate LXII. Fig. 29. Phleodictyon nodosum. X i. Fig. 30. Phleodictyon nodosum. ‘Transverse section of a fistula. 37. Fig. 31. Microciona prolifera. From a longitudinal section through a branch of the sponge shown in fig. 35. X 50. Fig. 32. Phleodictyon nodosum. Ectosomal skeleton of fistular wall as seen in a surface preparation. Beneath the fine dermal reticulum appear the fibers. X 21. Fig. 33. Microciona prolifera. From a transverse section through a branch of the sponge shown in Bigvsgen ers Fig. 34. Axinella acanthifera. Whole sponge. X 1. Pirate LXIII. Fig. 35. Microciona prolifera. Older, branched, form. X 1. Fig. 36. Microciona prolifera. Incrusting specimen, with lobes; on oyster shells. 1. Fig. 37 Acanthella corrugata. Whole sponge, viewed obliquely from above. X 1. Fig. 38. Axinella acanthifera. Transverse section of lobe showing axial skeleton and radiating fibers. X 15. Fig. 39. Axinella acanthifera. Longitudinal section through subcylindrical lobe, including both sur- faces. X 21. Pirate LXIV. Fig. 40. Pleraplysilla latens. Thick section through the sponge, vertical to the surface. Only the more superficial part of the sponge appears distinctly. Three fibers are seen to terminate in conuli. The transparent area between two of the fibers is a large canal extending radially inward from just below the surface. The large flagellated chambers are shown. X 50. Fig. 41. Hircinia ectofibrosa. Whole sponge. 2%. Fig. 42. Aplysilla longispina. Part of the surface. 114. Fig. 43. Hircinia ectofibrosa. Skeleton as seen in a macerated slice, vertical to the surface; principal fibers and connectives showing. The maceration was carried too far, and the superficial skeleton was injured. Xo. Fig. 44. Hirciniaectofibrosa. From asection radial to the surface, showing a principal fiber with attached connectives. X38. Pirate LXV. Fig. 45. Aplysilla longispina. Section vertical to the surface and through a conulus. A fiber is seen terminating in a conulus. The flagellated chambers are shown in the interior. ar. Fig. 46. Acanthella corrugata. Section vertical to the surface and radial to the margin; including margin and both surfaces of the sponge. X21. Fig. 47. Acanthella corrugata. From a section vertical to the surface and radial to the margin; below extreme marginal region. Mesial part of lamella and one surface are shown. X21. Fig. 48. Aplysilla longispina. Surface view of dermal membrane. Preparation photographed in water. X21. Pirate LXVI. [All figures reduced in reproduction to one-third original size.] Fig. 49. Spirastrella andrewsii. a, b, c, megascleres, 385; d, spirasters, 1,380. Fig. 50. Cliona celata. Tylostyles. 385. Fig. 51. Poterion atlantica. a, b,c, tylostyles. 385. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 52. 53- . Tetilla laminaris. Spicules a to g, X310; h, 800; a, stouter form of skeletal oxea; b, long 63. 64. SPONGES OF BEAUFORT (N. C.) HARBOR AND VICINITY. 179 Suberites undulatus. Tylostyles. 385. Stylotella heliophila. a,b,c, styles. 385. slender form of skeletal oxea; c, inequiended oxea of ectosomal brushes; e, stout protriane of lower part of body; /, protrizene from oscular margin; g, hair-like protrizene of general surface; h, sigmata. . Reniera tubifera. a, characteristic oxea and young stage; b, style; c, strongyle. 640. . Acanthella corrugata. a, style of the radial fibers, 138; b, oxeote modifications of the same, X138; c, d, e, megascleres with irregular ends, 640. . Microciona prolifera. a, b, skeletal styles, < 385; c, spinose styles, 385; d, isochela, X 1,380; e, toxe, X1,380. . Esperiopsis obliqua. Spicules a tod, X 640; e tof, X 1,380; a, smooth style; 6, spinose style; c, strongyle; d, slender tylostyle; e, toxa; f, twisted isochele. . Axinella acanthifera. a, style, common form; 6, style, stouter form; c, style, longer slendered form, projecting at surface; d, spinose style. 360. . Phoriospongia osburnensis. a, strongyles, 385; 6, tridentate isochele in side, dorsal, and ventral views, X1,380; c, sigmata, 1,380. . Hircina ectofibrosa. Ends of filaments. 1,380. . Lissodendoryx carolinensis. a, style, X640; b, tylote, 640; c, d, isochele in face and side views, X 1,380; e, sigmata, X 1,380. Phloeodictyon nodosum. Oxeas. 640. Aplysilla longispina. a, simple skeletal fiber, 138; 6, upper end of dendritic fiber, «138; c, dendritic skeletal fiber, including base of fiber and apex of one branch, 138; d, part of fiber, showing pith, < 640. SPiN, a Aas Re ‘ La ones sot? natalichaie a Pa Resin + pledgyathst dail foe fadolinle tohrneal salundl dah geahe _beeshartzo ot wile 4 ope ching? Aiea pearloid odor > yr RAS ‘pabi rsdn tue Livad ee doesitien! berated ree Pea deena yr soar ‘aie 0 Per seewol to ei vitor Pay Jour ye Rae a lek ry 6 atdeerngte oh Bt propihing= ee abner! fat prea r sayate rosie Py Boats Slit wilt jo utenmadiiow steone,.6 aries tathia Stele sigan 6h0X ober sofirgorat tre esvol renga by a) cd Btwn opeinttcti, ts, HBB atone here tstolvite ; 4.9 —.wedibord ; OBE TX | kot o b . ‘ SS \ ve ahyda BSE 1 AS >, Aa’ na) Mf gatiboty? pa ae BE yea rn fwd | \eead py fiat ts SBR i: ely ot raudols valve 6 sao eepiiaee pe tte ae pitony Ni as bd devi yaya -.' a i asinetberes pe Ka ramehin ht sentgio 9 to87. 175 ese cop peat.) awabeoleue ieteytoeloond’ AM. bp Mehns wegen be asatilateiods siqumie | 1) Joe bartuetal Siig Td fh tgyel ng hea Beg soo Pte aitthaton!naitiepie eilighash 4 e280 ig us ciwede 9dit wr ot. Fira we oie ta aur uals i Lie tie Deena ce-ealtey, SRO AN. ANRC 9, RRS sr Aeon Ot She Sebatar,. OME, OF he lea in ay heey table all ita web ler ae Bury. U.S. B. F., 1917-18. PLATE LVI. E LVII PLAT —18, if Buns Us o1 Bak. on a we, Sy Buu. U. S. B. F., 1917-18. PLATE LVIII. PLATE LIX. 1917-18. IShore st, WI Sp 18) 1G, PLATE LX. 1917-18. Buu. U.S. B. F., er PLATE LXI. Buu. U.S. B. F., 1917-18. PLATE LXII. Buy. U.S. B. F., 1917-18. Buu, U.S. B, F., 1917-18. PLATE LXIII. + LXIV. PLATE Buu. U.S. B. F., 1917-18. Buy. U. S. B. F., 1917-18. PLATE LXV. poe 4 Ww Rare Buy. U.S. B. F., 1917-18. PLATE LXVI. DRAGONFLIES AND DAMSELFLIES IN RELATION TO PONDFISH CULTURE, WITH A LIST OF THOSE FOUND NEAR FAIRPORT, IOWA z By Charles Branch Wilson, Ph. D. State Normal School, Department of Science, Westfield, Mass. ed Contribution from the U. S. Fisheries Biological Station, Fairport, lowa FOREWORD. The accompanying paper by Prof. Charles Branch Wilson, concerning a group of common insects in relation to fish culture, merits a special comment. ‘The author has not confined himself to a mere list of dragonflies and damselflies or to the recording of observations regarding their distribution, abundance, habits, and life history. He has supplied such necessary information, but, more important from the point of view of the Bureau of Fisheries, he has treated these insects fully and judiciously in their relations to fish, and thus in their relations to the food supply and welfare of man. We know that some insects, through destruction of crops and property or through injurious effect upon public health, are to be classed as insidious enemies of humanity and to be combated in every possible way; but there are others which we have learned to class as allies in the struggle for existence, since they make it possible for us to have useful articles of food and clothing, or are destructive to enemy insects. There are many insects of several orders, including the dragonflies and damselflies, which, before they begin to fly, spend a long period of existence in the water where they have direct or indirect relations to the useful fishes. The attitude assumed toward any of these must depend on whether they are found to be useful or harmful to fishes and to man. The relations of insects and fishes are complex. Voracious insect larve may destroy the fry of fishes or may consume food otherwise available to young fishes; they may destroy other and more dangerous enemies of fishes; or they may feed upon things that are not available to the desirable fishes and themselves become food for fishes. It is necessary to accumulate exact information and wisely to balance the good against the evil before we can determine whether the abundance of any particular aquatic insect should be opposed or promoted in the interests of an increased food supply from fishes. After a thorough analysis of all that has been known regarding the dragonflies and damselflies and all that has been learned in the course of this investigation, the author concludes with evident justification that these insects are, on the whole, of great eco- nomic importance, and he recommends them to the favor of the fish-culturist. Studies such as this, which can be applied not only to other insects but to various kinds of aquatic animals and plants, will necessarily have the effect of enabling us to apply more intelligence to the practices of fish culture and the production of food from private and public waters. H. M. Smiru, Commissioner of Fisheries. CONTENTS. & Page. ore word meaetmee ssiecemectemae cits cise ccsciiviss ss cnGesscsciuce-ducisaccetiecsctecessdeecccas © 182 MTD CLEO)T 16 Sob so doce COUR GEE UDB OD UE OeS MARCO RE See tae Se IS nee aaa eee ee 185 General description of the ponds and their environment....................00e cece eee eeeeeees 186 Aino anes oldraroniiies and: CamselMies,.«.c.6)<.210%,s.c1c= aiciois ow sig neb-tid nye wisi w owisicieisels vdeeee coeds 188 Relative abpndance Of different’ SPECIES <5.<./0/0.-.chs.cie sis e's «1s eyes clave Sod eveisle Guverdaie swears neve 189 ett cuinis tonya Ol atu ON Alen. ie eee tt cis dn ace areca orci ctetne icp ernie cin, Leaps Seats watt giac Ss s.eeivmeleceres 192 Moree atts OOOO Ate TV INPNSa te mctte eaN Acie t oi eroninaimee rahe: Seba s paeetide ns wae aeons 193 Marth patts ohodonate tnavOsse toils ond Sick aciar eaten) cise nis wb bid Measles Nes wks Farah eles ods 199 Heonomic relations between odonatesiand’ fishes. -- - 2.2.5. -2-2ecece- ns scene cede sewsn- ences 200 Hane GiOdOnAL EMMY IN PHShen eae crate Sens = hn oh oa sae anis be aie ees se colon ection, Siem aece 201 PES TUG TALES CONIC OER ALE ALY AIMED INS whet Mees cVe yoo hsp cua) ctaccrs ocayal> ane bis oye sintcbercnete whose ois sfeie esarep ciate acess, zusroraie art 209 Hane lof adoriate tiMAagOs: octets cpa) she 2 ois - aw sas Eohemis sim? cover 15 to 25 mm. ‘This, in addition to the covering the lower part of the faceuptothe lurching forward of the body, enables them to epbesistee grees catch insects like Corixa, much more agile than themselves, and even to capture small fishes. To assist in holding their prey, the lateral lobes of the mask are toothed along their inner margins in the A’schnide (fig. 11). In the Libellulide they are armed with a long, slender spine at the tip and a row or raptorial sete behind this along the outer margin, varying in number in the different species. There is a crescent of similar sete, also varying in number, along the body of the mask, called the mentum, on either side of the median line. And there are more or less regularly arranged spines and hairs along the remaining margins of all three lobes. Figures 12 to 14 illustrate some differences between the species. Maxill@.—Once grasped between the lobes of the mask, the prey is drawn quickly to the mouth, where there are two pairs of organs ready to dispose of it. The first of these, the outside pair, are the maxille, which are very much alike in all nymphs; each maxilla has two fingerlike branches or rami, the outer (ventral) of which is DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 195 Fic. 7.—Side view of head of nymph of Libellula luctuosa, mask folded. Fic. 8.—Same, mask extended. Fic. 9.—Side view of head of nymph of Anaz junius, mask folded, Fic. 1o.—Same, mask extended. Fic. r1.—Mask of Anaxjunius. Fic. 12.— Mask of Libellula luctuosa. Fic. 13.—Mask of Erythemis simplicicollis. Fic. 14.—Mask of Pachydiplax longipennis. 196 BULLETIN OF THE BUREAU OF FISHERIES. armed with stout, curved claws, while the inner is covered with stiff hairs; these maxille are evidently used to help hold their prey securely (figs. 15 and 16). Mandibles.—The second pair are the mandibles; they are much stouter, are hard and chitinous, and are armed with strong teeth (figs. 17 and 18). They can easily crush the shells of small pond snails like Limnea, Physa, and Planorbis, or they can bite through the hard chitin covering of beetles and water boatmen. ‘The food con- tents of the stomachs of all the nymphs examined shows that the mandibles are used chiefly for crushing the food and not for chewing it. It is chewed only enough to get it down the gullet, and much of it is swallowed whole. Gizzard teeth—The real mastication takes place in the gizzard, and for this purpose the wall of the gizzard at the posterior end is armed with four longitudinal ridges of chitin—two dorsal and two ventral. Each ridge carries projecting teeth, whose number and arrangement varies a little in different species. The general character of these Ac. 15.—Maxilla of Anax junius nymph. Fic. 16.—Maxilla of Erythemis simplicicollis nymph. iG. 17.—Mandible of Anaz juniusnymph. Fic. 18.—Mandible of Erythemis simplicicollis nymph. toothed ridges is well shown in figures 19 to 24. The churning of the gizzard grinds the food against the teeth and soon reduces it to finer fragments; it then passes on into the intestine. DAMSELFLY Nympus.—The structure of the mouth of the damselfly nymph is in all respects similar to that of the dragonfly. The mask (fig. 25) is more like that of the Libellulide, with raptorial sete on the lateral lobes and the mentum, but the lateral lobes only cover a very small portion of the lower face. The mandibles (fig. 26) and the maxille (fig. 27) are so much like the larger ones of the dragonflies that they can be recognized at once by comparison. In the gizzard we find a somewhat different arrange- ment; instead of 4 chitin ridges there is some multiple of 4 up to as many as 32, 8 and 16 being the most common numbers. Each ridge has a row of small spinelike teeth along the anterior half of both lateral margins; there is a narrow space through the center which is unarmed, and the whole posterior surface is covered with short stout spines, curved forward (fig. 28). Such a mill ought to be able to grind the food into very smail fragments, and we find that this is actually done. In other damselflies the gizzard varies greatly both in the number of ridges and in the size and number of the teeth. The differ- ences in the various genera and species have been admirably worked out and figured by Miss Higgins (1901). DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 197 Wes. 19 and 20.—Dorsal and ventral tooth from gizzard of Anaz junius nymph. Fics. 21 and 22.—Dorsal and ventral tooth from gizzard of Libellula luctuosa nymph. Fics. 23 and 24.—Dorsal and ventral tooth from gizzard of Epicordulia princeps nymph. Posterior end of each tooth toward the right. Fics. 25 to 28.—Enallagma nymph: 2s, mask; 26, mandible; 27, maxilla; 28, gizzard teeth, the upper edge of figure anterior. BUPLETIN OF THE BUREAU OF FISHERIES. 198 % ! yooy ajqeAo om ‘ous tuIn}uoUr JO aqo] UvIpant ‘7ue SyseUT JO aqoy [e19}¥] ‘77 ‘siejour ‘aiqrpueur UM ‘SIOSIOUI * ‘pe leyprxeur a 33 ut ‘yooy pus a[qeAonrur ‘a *QAOQE SE QOULIFIUSIS SUIYS Sia}}a] 4oTJO SySeur fyseur ‘cf stpostayduus stumayjdiy yo odvuy—'vf 07 ‘ajqrpueur ‘If Seyjixeur ‘of tysvru ‘6c -sntwnl xvuy jo OdvuIy—"1f 0} 62 “soy 1 S| org ‘dyed ‘¢ ‘ourenbs ‘sw DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 199 MOUTH PARTS OF ODONATE IMAGOS. DRAGONFLY ImMacos.—When the nymph transforms into an imago, the hinged mask disappears, and the labium or lower lip is now attached directly to the ventral surface of the head. In consequence there is a much greater similarity in the mouth parts of the im- agos, though there are still some differences. In general, the labium of the imago consists of a basal portion, corresponding to the mentum of the nymphal mask, but destitute of raptorial setee and very much shortened. From its anterior margin projects the median lobe, greatly enlarged in Anax and Aischna (fig. 29) and covered with hairs, but reduced to a small triangular piece in the Libellulids (fig. 32) and more or less free from hairs. On either side of the mentum is a side piece known as the squame (ms) which sup- ports the lateral lobe. These latter now fold together across the front of the face in all the imagos. In Anax and Atschna they are enlarged into concave lamelle, strongly convex on their outer borders and covered with hairs. The inner border ends distally in a sharp end hook (e) which is immovable; just outside of this is a larger movable hook (mo), which is rounded and palplike and covered with hairs. When the lobes are folded across the front of the face, these four hooks meet on the median line, but the margins of the lobes beyond them diverge rapidly. In Erythemis (fig. 32) and the other Libellulids the margins of the lateral lobes, when folded, meet each other in a long median Bee ea ere iors mtn, emesis zy ™™ line, both hooks have practically entirely disappeared, and the lobes are covered with hairs. The mandibles (figs. 31 and 34) and the maxille (figs. 30 and 33) have changed a little in detail, but are practically the same as before. The teeth of the mandibles are very strong and may be divided into two sets, the incisors (in) at the tip of the mandible, which are long, curved, and sharply pointed, and the molars (m) near the base, which are much shorter and armed with separate cusps or points. The maxille still retain the outer lobe or palp (p), which is curved and covered with hairs, and the inner lobe, which is armed with curved and sharply pointed teeth and a pad covered with long sensory hairs. DaAMSELFLY ImMaGos.—In the damselflies the general structure of the mouth parts is the same as in the dragon flies. Here the median lobe of the labium (ml, fig. 35) is fully as long as the lateral lobes, is divided by a deep median fissure, and is covered with long hairs. The lateral lobes retain the immovable end hook (e), which is very long and slender and curved to an acuminate point, and the movable hook (mo), which is also long and narrow, but is bluntly rounded and covered with hairs. These lobes are relatively much narrower than those of the dragonflies, and, when folded, only the terminal hooks meet on the median line. The mandibles (fig. 36) and maxille (fig. 37) are similar in all respects to those of the dragonflies, except that there is a sharper distinction between the incisors and the molar in the mandibles, while the maxillz have long hairs on the outer margin near the base. The gizzard in the imago is relatively much smaller and weaker in the nymph, and has very little functional use. The chitin ridges or folds along its walls are still retained, 200 BULLETIN OF THE BUREAU OF FISHERIES. but the teeth are either completely lost, as in the dragonflies and in Lestes among the damselflies, or they are reduced to a much simpler form. The imago evidently chews its food before swallowing it, as we may well believe after watching one munch its prey. This brief description of the mouth parts of the nymphs and imagos will enable us to understand better both how they secure their prey and how they dispose of it afterwards. ECONOMIC RELATIONS BETWEEN ODONATES AND FISHES. The artificial propagation of fish falls very naturally into three great divisions: 1. Suitable methods of obtaining and hatching the eggs; 2. Care and protection of the young after they are hatched; 3. Provision of an abundance of the right kind of food. Our national and State fish hatcheries are concerned very largely with the first two of these, and the progress they have achieved is marvelous, considering the difficulties surmounted. Ina comparatively brief period of years they have accumulated a wealth of accurate information and statistics, which have attracted the attention and awakened the admiration of the entire world. But the last factor, in so far as it concerns pond fishes, has thus far received almost no attention in this country. Europe has been studying the problems connected with fishponds for many years and has far outstripped us along these lines. In fact, we have made hardly a beginning as yet, and the few facts that have been ascertained still lack correlation and logical arrangement. In speaking of the food relations of insects and fishes Needham (1901, p. 395) said: “And so little are the essential features of good foraging ground understood that each planting of fry in a new place is still largely an experiment. * * * Any new study of fish food should include the study of the feeding grounds, feeding habits, choice of food offered, and conditions that make for the continuance and possible increase of the food supply.’’ In spite of the 15 years that have elapsed since then, the statement retains practically its full value to-day. Prof. S. A. Forbes and his associates in Illinois were pioneers in this aspect of fish culture, and have put out much valuable information on the food of fresh-water fishes during the last 35 years; an admirable summary will be found in Forbes, 18885. More recently papers have been published by L. L. Dyche, 1914; Wm. E. Meehan, 1913; A.S. Pearse, 1915; Geo. C. Embody, 1915, and Dr. Robert E. Coker, 1915. These papers are all excellent in both their subject and its treatment, but of necessity they are general in character and do not treat any of the phases in detail. For this reason and for many others it is believed that a summary of the economic relations between odonates and fishes will prove of interest to all who are, and especially to those who may become, engaged in pondfish culture. At present the old idea prevails that the nymphs of the Odonata, especially those of Anax, A!schna, and other large species, are very destructive to small fishes. So far as known, the other side of the question has never been presented, and we have not only drawn a one-sided and biased conclusion, but we have also been led into the common error of condemning the many for the sins of the few. A hawk steals a farmer’s chicken, and immediately all hawks are condemned as pests and robbers, irrespective of species, and a loaded shotgun is kept for their reception. Similarly because Anax and A’schna nymphs have been known to kill small fish, all dragonfly nymphs have been condemned as nuisances and dangerous to have around a fishpond. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 201 Such a conclusion is, of course, not warranted, and even with reference to Hschna itself it does not follow that it must be considered dangerous simply because it eats a few small fish. Needham (1903, p. 213), one of our most experienced observers, said with reference to 4schna constricta, which was common at Bone Pond, one of the three propagating ponds controlled by the Adirondack hatchery in New York, and artificially stocked with brook trout: “I have not been able to determine as yet whether in rela- tion to trout culture Aischna is more disadvantageous than otherwise. It eats a few fry and it eats the food of the larger trout, but, on the other hand, it furnishes a moderate supply of food itself for the larger trout.” In order to solve such a problem satisfactorily we must consider in detail all phases of the economic relations between odonates and fishes. An endeavor to do this has been made in the following pages by a careful con- sideration of the food and enemies of dragonfly and damselfly nymphs and imagos, by asummary of all that is known of nymphs and imagos as fish food, and by extensive observation and experimenting. FOOD OF ODONATE NYMPHS. The alimentary canals of 50 nymphs of each of the four dragonflies enumerated in the following table and 50 others of various damselfly species, making 250 in all, were examined to ascertain what food had been eaten. The presence in a nymph’s digestive tract of any of the various kinds of food listed was represented by a single unit, whether found in large or small quantities. The figures in the table, therefore, tell us how many of the nymphs had eaten any specified food, but do not indicate how much they had consumed. Foop oF 250 ODONATE NyMPHS FROM Farrport, Iowa. [The figures indicate the number of nymphs in which each kind of food was found.] a a . | a 4/3 2/3 f=} ° | h.e| a ./-3)ala g| ale) 4 i¢/ S| 8/4 6| &| 8 $|a;}g] & : #/8)2/ 8 : S| e/a) 8 Contents of alimentary canal. 3 3 t g a Contents of alimentary canal. 3 | 3 ‘3B a a s)a/E\2/s 5/4] 2/2|3 Fle la] se) 8 Pisla]o13 o/yr] a W)ole Tsai g E 4/218 g g/2/ 213) 42 alalala Be | gai/ala i Mollusca: Crustacea—Continued. | Physa shells, mostly fragments... .. 16 | 6|16] 4 8 Simocephalus, sp s| 2 4 8 Planorbis shells, mostly fragments..] r5 | 34 | 22] 4 | 12 Daphnia, sp........ a [adn 4| 17 Beetles: Pleuroxus, sp 03. | Seal gis 5 Miytiscns Narver eyes 2s tle. winks Zo,] 8}. a}, a}... Copepods, fragments unidentified...| 2) 2] 4| 14] 15 Haliplid beetle, elytra of adults..... Bg [eS ler Lek Diaptomts,/spe 2. AE ht ere aS ae a | 7 6 Peltodyes larve, jointed hairs...... crear, 10,6, | fo Seles ood Hanletiabaisst cs sovavecckedus «tee Boe Wee ty ree Mees Pe Diptera: CrayGSh; Spi eae. eran ot Py dhl aed eek a ale Bs Ceratopogon ay Be ey: See 2| 3| 6|....| 2 || Odonata: Chironomid larve. at he bie 8 Damselfly nymphs, unidentified....| rr TS eyes 2 Mosquito larve.. Or he a] PAs 6 |! Ischnura verticalis nymphs......... 12,103.02] Loge Sammmndriarye. 32 ee! Peete ae eB Naa 9 Enallagma, sp., nymphs ........... 3) 2] 5] zl]. Hemiptera; Corixa, sp.,elytraofadult..|14| 4] 3] 1/|....- Lestes, sp., mymmphs,.............+5 Oth wail Saul sabes: Sialidse: Sialid larve................... era ag iPor isa Libellula juctuosa, nymphs........ Bh Piodd lacie gas WH. Perlide; Stonefly larve.............05. eee ie nee) ae ee Anax junius, nymphs.............. Bloc al Wendel ee ie: Ephemeride: Mayfly larve............ 27 7| 4 6 Erythemis simplicicollis, nymphs..| 6| 4] 1 |....|.... Crustacea: Alge: Entomostraca, fragments unidenti | pesmids 0! +. MAC cathe het ee 2 9/ 4] 2] 33 Shor Poe te Se eee 5 | 21 | 26] 15] 20 Diatomsis.. 8. aioe cee te aeeeetgee (ane! amet BIE aalle= Cladoceran eggs, withephippium...| 2/ 7| 9] 6) 10 Sponge spicules, in masses. 7H oct: liee! ay eed Geet ces Bosmina, sp, 3} 8} 6] 4]. 12 CEdogonium, scattered filaments....|....]..2.]'"6 3) 14 Cypris, sp... 6/35] 2|'20) 15 Alge filaments: 2.0.0.0. 0... 0200s 7| 7|32| x1 25 202 BULLETIN OF THE BUREAU OF FISHERIES. GENERAL CHARACTER OF THE Diet.—On the whole there is a remarkable similarity in the diet of the various species. Eleven of the food items appear in all five of the columns and nine more are found in four out of the five. There are only three foods confined to a single species and three others that are restricted to two columns. Perhaps the most noticeable items are those which begin and end the list. Two species of snails are very common in all the ponds, and upon these every kind of nymph examined had been feeding freely. Not only had more than half of the nymphs partaken of these snails, but in several instances no other food was found in their digestive tract. With reference to the algce, it is of course understood that inasmuch as the nymphs catch their prey among the alge, they would be expected to swallow some of the latter. Hence its presence was not noted unless in sufficient quantity to make it reasonably certain that it had been taken voluntarily. Like the snails, in a few instances it con- stituted the sole article of diet. The other popular foods were the mayfly larve and the small crustacea, the latter being consumed in large quantities. Cannibalistic tendencies are shown by the presence of odonate nymphs in all five columns; and in four out of the five, nymphs were found which had eaten others of their own species. There is a good showing of beetle larve and adults, and of adult water boatmen, all of which are injurious to young fishes. SPECIFIC DIFFERENCES IN THE Diet.—Notwithstanding the wonderful agreement just mentioned, a careful examination reveals also striking differences in diet. Every one of the Anax nymphs was fully grown, with well-developed wing cases, and was captured in pond 4, which contained only adult buffalofish. Of course the nymphs could not eat these fish, and hence the absence of fish in their diet is a matter of necessity rather than choice. (See p. 206.) Neither did the fish eat the nymphs, however, and this probably accounts for the exceptional abundance of the latter, as evidenced by the exuvize obtained. The bulk of the food of these nymphs was made up of mayfly larve and snails, but it is worthy of note that they also ate large quantities of Dytiscus larve, the water boatman Corixa, and small crayfish. The proportion of odonate nymphs in their food was much greater than for any of the other species examined. One very large specimen had eaten nothing but crayfish, and its stomach was packed full of their shells and claws, which had turned red in color like a boiled lobster. The Dytiscus larve were identified by their heads and mandibles, Corixa by the peculiar color pattern of the elytra, and the Haliplid beetles also by the color pattern of the elytra. All three of these foods were especially abundant in this pond. The bulk of the food of the Libellula nymphs consisted of snails and various entomostraca, Cypris being a particularly toothsome tidbit; there were also quite a number of damselfly nymphs and a good representation of the beetle larve. Fios. 41 to 43.—Development of Libellula pulchella; 41, egg; 42, mask of newly hatched nymph; 43, newly hatched nymph. whitish. The whole body is transparent and pale white, a little yellowish on the head and at the posterior end of the abdomen; the circulatory system is deeper white. The Mask.—Mentum three-quarters wider than long, distal margin two and a half times the length of the proximal; four sete on the inner surface, near the lateral margin, one proximal to them on the margin itself; three mental setz on either side in nearly a straight line; two tiny spines at the center, near the distal margin. Lateral lobes one-fourth longer than wide; two sete on the outer margin, near the base; two setz on each blade; raptorial seta longer than the movable hook, the latter long and narrow; distal margin with six teeth near the outer border, the first two the longest. HYBRID BETWEEN LIBELLULA PULCHELLA AND L. LUCTUOSA. A male L. pulchella and a female L. Juctuosa were captured mating on the shore of pond 4D July 9, 1917. The female deposited about roo eggs in a tumbler of water, and these subsequently hatched on July 20. The Nymph.—This nymph is unlike those of either pulchella or luctuosa, especially in its markings. The head is one-third wider than long, and the same width as the abdomen, grayish yellow; the eyes are sulphur yellow, with smal! black retinal dots; the rest of the body is pale yellowish, but without much color and quite transparent, except along the lateral margins of the abdomen, where each segment is dark gray anteriorly and white posteriorly. The whole body is quite hairy, with a row of stout DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 241 bristles along the dorsal surface of the abdomen, between the réspiratory system and the lateral margin on either side. The antenne are relatively large; base white with a pale grayish distal band; first joint white through the center, pale grayish at either end; second joint with proximal three-fifths dark and distal two-fifths white; third joint dark, tipped with white. These nymphs did not ascend to the surface of the water, like those of luctuosa, but stayed down near the bottom of the tumbler. The Mask.—Mentum twice as wide as long; the distal bor- der two and a half times the proximal; the lateral margins curved; three sete on either side, near the lateral margin; three mental sete in a straight line; a slight lobed promi- Wis. 44 and 4s—Hybrid from male Libellula pulchella and female Libellula luctuosa; nence behind the center of the 44, mask of newly hatched hybrid nymph; 4s, newly hatched hybrid nymph. distal margin, without spines. Lateral lobes one-third longer than wide; two sete on the outer border near the base; two on each blade; raptorial seta considerably longer than the movable hook, the latter short and slender; distal margin with four teeth near the outer edge, the first two the longest. ERYTHEMIS SIMPLICICOLLIS. The first females were seen laying in pond 4D on July 12, 1917. Eggs laid July 13 hatched July 23. The Egg.—The egg of this species is an elongated ellipse, with rather pointed ends, and is a little more than twice as long as wide, the respective diameters being 0.60 mm. and 0.27 mm. ‘The neck has a base narrower than its height, divided at the tip; there is a small process on the posterior end of the egg; the yolk granules are minute. The Nymph.—This nymph has the most pigment of any of those examined, and is banded brown and yellowish white; the eyes are light orange yellow with comparatively large black spots; the sides of the head are brown shading into light yellow on the median line; each thorax segment is brown 0.5 mm. 46 Figs. 46 to 48.—Development of Ervtkemis simplicicollis: 46, egg; 47, mask of newly hatched nymph; 48, newly hatched nymph, bordered with dark yellow along the lateral margins and light yellow in the center, thus leaving a clear yellow longitudinal streak through the center of the body; the ninth and tenth segments are yellow; the legs are dark proximally, with a white band across the distal ends of the coxez, femora, and tibiz. The central yellow line of the head passes over the forehead and down between the antennz onto the labrum. On the back of the head it is widened considerably and usually mins out into two rounded points on either side, with two small, brown spots, one on either side of the midline, at the level of the posterior points. ‘Through the thorax it is narrow, then widens again on the first three or four abdomen 242 BULLETIN OF THE BUREAU OF FISHERIES. segments, and fades into small, yellow dots on the anterior margins of the fifth, sixth, and seventh, and sometimes the eighth, segment, disappearing entirely on the ninth, and often on the eighth segment. The dorsal appendages on the tenth segment are light brown at the base, yellow at the tip; the lateral and inferior appendages are dark brown. The base of the antennz is yellow distally and proximally with a narrow, black band through the center; first joint dark, with a narrow, distal, white band; second joint with a broader, distal, white band; third joint entirely blackish brown. The Mask.—Mentum with the width to the length as 5 to 3, distal margin two-thirds wider than the proximal; three lateral sete, one marginal seta; three mental sete in a straight line on either side of the center. Lateral lobe with its length to its width as 9 to 7; two marginal sete on the outer border; two setz on the blade of the lobe; raptorial seta just reaching the tip of the movable hook, the latter long and stout; distal margin crenate, without teeth. LEUCORRHINIA INTACTA. The sexes remain in union only a short time and usually alight on some convenient weed or bush near the water’s edge, where they can be easily secured. When thus captured, the female is ready to lay as soon as taken from the net and will deposit 100 to 200 eggs, the first 50 or 75 coming in a mass stuck together, the others coming singly. Eggs laid July 18 hatched July 3o. The Egg.—This egg is more nearly spherical than that of Libellula luctuosa, and has diameters of 0.60 mm. and 0.40 mm., respectively. The neck is wider at the base than it is high, and there isa distinct segmentation at about the center; the tip is not divided, and there is no process on the posterior end of egg. The yolk granules are comparatively large, and the jelly envelope is ragged around the surface of the egg and not smooth, as in other species. The Nymph.—Although the imago of Leucorrhinia is much smaller than that of Libellula luctuosa the 49 Fics. 49 to 51.—Development of Leucorrhinia intacta: 49, egg; so, mask of newly hatched nymph; sr, newly hatched nymph. nymph is considerably larger, measuring 1.20 mm. in length; the head is one-third wider than the abdo- men; the front legs are the same length as the body, the others proportionally longer, and all three pairs slender. Both the body and the legs are quite transparent and pale orange yellow in color, the legs and antenne transversely banded with gray and white, the eyes darker orange yellow, the black spots very small. The sides of the head and a wide transverse band across the anterior portion of each abdomen segment are whitish, the posterior margins of the segments are tinged with brown. The bases of the antennz are whitish yellow proximally, with a narrow, gray, distal band; first joint entirely white, a little grayish through the center; second joint with a very narrow proximal band and a wider distal band white, grayish through the center; third joint entirely gray. The Mask.—Mentum three times as wide as long; the distal margin twice the length of the proximal; no lateral sete; two marginal sete; three mental sete on either side, not in line, but the proximal one nearest the midline; two stout spines at the center of the distal margin. Lateral lobes with length and width in the proportion of 13 to 9; two sete on the outer margin on either side; two sete on the blade of each lobe; raptorial seta reaching considerably beyond the tip of the movable hook, the latter long and narrow; three small teeth near the outer edge of the distal margin, the central tooth the largest, the rest of the margin smooth. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 243 PLATHEMIS LYDIA. The females of this species remain at some distance from the water, except at the time of ovipositing; one caught a quarter of a mile from the ponds deposited eggs freely on July 25, which began to hatch August 2, and were all hatched by August 9. The Egg.—tThese eggs are elongate elliptical and rather pointed at both ends, just twice as long as wide, the two diameters being respectively 0.60 mm.ando.30omm. ‘The neck isas wide at the base as it ishigh, with a distinct segmentation near the tip, the latter being undivided; there is no process on the posterior end of the egg. The Nymph.—The general color of the nymph is yellowish white, with light-brown markings. On the center of the dorsal surface of the head is a longitudinal, club-shaped, brown area and on either side a broken semicircle; there is a large, irregular, brown spot in the center of the posterior thorax and anterior abdomen, and another smaller spot over the posterior end of the respiratory system. The eyes are dark brown, with black, retinal spots; the tracheal system is yellowish brown. The head is a trifle wider than long and not narrowed behind the eyes; the thorax is two-thirds as wide as the head; the 62 Ye 0.5 mm. ey Fics. 52 to s4.—Development of Plathemis lydia: 52, egg; 53, mask of newly hatched nymph; 54, newly hatched nymph. abdomen the same width as the head, sharply pointed posteriorly, with strongly convex, lateral margins and no color markings except the two spots already noted. The base and basal joint of the antenne are whitish; proximal end of second joint and the whole of the third joint light gray, distal part of second joint white. The legs are light gray and white; the tips of the posterior processes of the abdomen are dark brown; the terga of the thoracic segments are light gray. The Mask.—Mentum twice as wide as long, with convex sides; distal margin two and a half times the proximal; one lateral seta near the base on each side and two marginal sete; three mental sete on either side of the center in a curve concave to the midline; two small spines at the center of the distal margin. Lateral lobes one-half longer than wide; two sete on the outer border, none on the blade; raptorial seta just reaching the tip of the movable hook, the latter short and stout; distal margin with seven teeth reaching nearly the entire width, the third tooth the longest. EPICORDULIA PRINCEPS. The eggs of this species are laid in long, jelly-like strings, similar to those of Tetragoneuria. Some of these strings were obtained from the leaves of Potamogeton illinoiensis in pond 3D on August 9, 1917; but, of course, there was no way to tell when they were laid; they began to hatch within a day or two. The Egg.—This egg is the largest of any here described, being 0.72 mm. long and 0.40 mm. wide. The neck is a minute process of the same height and width, without segmentation, and there is no process on the posterior end of the egg. 244 BULLETIN OF THE BUREAU OF FISHERIES. The Nymph.—Like the egg from which it was hatched, the nymph is a little larger than any of the others here included, being 1.25 mm. long and 0.42 mm. in diameter. Its general color is white, the eyes and the center of the abdomen reddish brown, the respiratory system bright yellow. There are two com- paratively large horns on the dorsal surface of the head at the posterior margin, with their tips turned back like hooks; that is, they are slightly “‘cultriform’’ (Needham). The inferior posterior appendages of the tenth abdomen segment are large and are turned over vertically at right angles to the body axis. The 55 Fics. 55 to 57.—Development of Epicordulia princeps: 55, egg; 56, mask of newly hatched nymph; 57, newly hatched nymph. antenne are banded with black and white in sharp contrast; the base with a narrow, white, distal band; the basal half of the first joint black, the distal half white; the second joint the same, the third joint entirely white. The legs are transparent, with very little differentiation in color, the tips of the basal joint, the femur, and the tibia being somewhat whiter than the rest of the joint. The thorax is pale white, without any pigment; the abdomen is brownish red through the center over the intestine, with short, dark spots on the posterior margin at the outer edge of each segment, the rest white. The Mask.—Mentum three-fifths wider than long, its sides slightly concave; the distal margin a little less than twice the proximal; one lateral seta, two marginal sete on either side; three mental sete on each half in a straight line, the distal one nearest the midline; four stout, sharp spines at the center of the distal margin. Lateral lobes about the same length and width; two sete on the outer border, none on the blade of the lobe; raptorial seta reaching beyond the tip of the movable hook, the latter short and stout; distal margin with 10 teeth covering its entire width, the outer 5 much larger than the inner 5 and cultriform. Lateral seta on mentum often lacking as in figure 56. ENALLAGMA HAGENIT. The eggs were laid in the leaves of crex grass; more than a thousand were obtained from five grass leaves July 26, 1917. These began to hatch in two weeks and continued hatching for ro days, but, like the eggs of Epicordulia, there was no way to determine just when they were laid, so that the period of incubation is uncertain. The Egg.—These eggs were in the form of an elongated ellipsoid, the long diameter four times the shorter one, the anterior end broadly and bluntly rounded, the posterior end pointed; neck short and broad and brown in color. The Nymph.—The head is transversely elliptical in outline, the two diameters in the ratio of 8 to 5; there is a pair of sete just inside of each eye on the dorsal surface, another pair close to the mid- line, just behind the anterior margin, and a single seta at the center of the margin itself. The thorax is considerably narrower, but almost twice as long as the head; the legs are long and stout, the posterior pair reaching somewhat beyond the center of the caudal gills. The abdomen isa little wider than the thorax anteriorly and does not narrow much in front of the seventh segment. The antennz are long and stout, with a gray band at the base of each joint; caudal gills as long as the rest of the body and DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 245 cylindrical, covered with short and straight spines or sete. All the leg joints, including the coxe, are sparsely armed with sete; and there is a row of sete also along the dorsal surface of the abdomen close to the lateral margin. The color is a uniform, creamy white, except the three dark bands on the antenne and a narrow band on each femur and tibia of the legs. The Mask.—Mentum triangular, one-half wider than long; distal margin twice the length of the proxi- mal; one mental seta on either side, one lateral seta at the base of each lobe, two marginal sete; a row of nine small spines just behind the distal margin. Lateral lobes twice as long as wide, with one marginal seta and two raptorial setae; movable hook long and stout; an accessory spine just outside the base of the distal raptorial seta. Nis (A |>==4 Si K NS & Sas SS i 1 LEZ IIA | ESR, TI ns Y, A OR \y (y PSO rs Sh TTT pipe 453) Sania VA it SF ZA SS 58 Fics. <8 to 60.—Development of Enallagma hageni: 58, egg; 59, mask of newly hatched nymph; 60, newly hatched nymph. ENALLAGMA SIGNATUM. These eggs were obtained from pond-lily leaves on July 28, r917, and began to hatch 18 days later, the period of incubation being again unknown. The Eqg.—The eggs are considerably like those of E. hageni, but they were arranged very differ- ently in the tissues of the leaf. Those of hageni were inserted without any definite order anywhere in the leaf; these were arranged in a semicircle on the underside of the leaf around some convenient hole or close to the margin. The female evidently thrust her abdomen down through the hole or down over the edge of the leaf and, reaching as far as she could, inserted the eggs into the leaf. The Nymph.—The nymph is just 1 mm. long, exclusive of the caudal gills; head wider than long, the two diameters in the proportion of 7 to 5, the anterior and posterior margins both strongly convex; thorax shorter than the head and two-thirds as wide; abdomen the same width as the thorax and taper- ing gradually backward; caudal gills short and slender; legs also comparatively short and slender, the posterior pair not reaching the center of the caudal gills; claws short and stout. The antennez are long and stout; the eyes at the center of the lateral margins project strongly. There are no sete on the dorsal surface of the head, on the thorax, or on the coxe of the legs; there is a single seta on the center of each abdominal segment near the lateral margin on each side; and the setz on the legs and 246 BULLETIN OF THE BUREAU OF FISHERIES. caudal gills are small and scattering. The ground color is white, covered with a complicated pattern of light russet brown; last four abdominal segments, legs, and caudal gills nearly all brown, with narrow stripes of white; respiratory trachee bright golden yellow. The Mask.—Mentum one-half wider than long, with slightly convex sides; distal margin nearly three times the proximal; no lateral sete, one marginal seta near the distal end, one mental seta; distal margin smooth. Lateral lobe three-fourths longer than wide; no sete on the outer margin; raptorial seta just reaching the tip of the movable hook, the latter long and stout; a minute accessory spine out: side the base of each raptorial seta; distal margin deeply toothed, inner tooth cultriform. 61 Fics. 61 to 63.—Development of Enallagma signatum: 61, egg; 62, mask of newly hatched nymph; 63, newly hatched nymph. GENERAL CONCLUSIONS. 1. Odonate nymphs feed upon small mollusks, insect larve (including smaller nymphs), pupz and adults, entomostraca and larger crustacea, and alge. Some of their food, such as Chironomid larve, mayfiy larve, entomostraca, etc., is the same as that of young fish, but they also eat the larve or adults of many animals that are directly harmful to small fishes, such as diving beetles, water boatmen, crayfish, and Cypris. 2. A few of the largest species may sometimes eat a small fish under natural con- ditions, but this is apparently due to stress of hunger and the lack of other food. War- ren has proved (p. 206) once for all that the diet of a nymph in captivity furnishes no criterion whatever as to its natural food. Careful observations under natural condi- tions show that even an Anax nymph need not be regarded as a menace to fish culture, but that it may become actually beneficial. 3. Odonate nymphs furnish one of the very best foods for fishes; the small species and the young of the larger species are freely eaten by the fingerlings of practically all DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 247 our fresh-water game and food fish, while the larger species when fully grown are just as toothsome to adult fish. The large percentage of nymphs in the food of fishes from the Fairport ponds (table, p. 228) effectively answers the objection that odonates rob the fishes of a part of their food. 4. If the Anax, A’schna, and other large nymphs are preying upon the fish in any given fishpond, this should be interpreted by the fish-culturist as evidence that the supply of food in that pond is running low. His efforts can better be directed toward replenishing the food supply than toward getting rid of the nymphs. On the other hand, if they do not disturb the young fish the food supply is adequate, and they them- selves will contribute to it in due time. They thus furnish a convenient means of testing the food supply, since it is an easy matter to examine their stomachs and find out what they are eating. 5. Odonate imagos feed exclusively upon adult insects; their prey, like that of the nymphs, sometimes includes insects that are beneficial, such as other odonates, honey- bees, and hymenopterous parasites; but the bulk of their food is made up of insects that are either positively injurious or negatively of no practical importance. Among these may be mentioned gall flies, tsetse flies, plant lice, leaf hoppers, ants, and all kinds of small moths and butterflies. They also confer an inestimable benefit upon mankind by waging an incessant warfare upon house flies, mosquitoes, black flies, and gnats. This one benefit alone far outweighs any harm they may do to the fish and should earn for them a cordial welcome to every fishpona, present and future. 6. Odonate imagos, like the nymphs, furnish good food for adult fishes, as is evi- denced by finding them in the stomachs of various fishes taken under natural condi- tions and from the fishponds. It has also been demonstrated by feeding live dragonfly imagos to game fish in fishponds. 7- Both nymphs and imagos are important factors in establishing a natural equi- librium in the fauna and flora of the fishpond and its immediate environment. Other things being equal, such an equilibrium contributes materially to success in fish culture, and it can not be obtained without the presence of the nymphs and imagos. Inci- dentally, if properly chosen, the imagos will add greatly to the attractiveness of the fish- pond and its surroundings. 8. Hence care should be taken that the pond is supplied with odonates, as well as with fish; after the original stocking they can be left to take care of themselves. For stocking purposes use common local species of dragonflies and damselflies; Anax, L. pulchella, Plathemis, and the damselfly genera, Argia, Enallagma, and Lestes are so cosmopolitan that they will make good stock almost anywhere in the United States. Here again the odonates constitute a sort of visible pulse of the life of the pond; so long as they remain vigorous and healthy the pond life is probably moving along smoothly. g. Small breeding ponds along the margin of the fishpond from which the fish are excluded will materially increase the supply of all aquatic insects, including the odonates. If these are started in the spring or fall, the best odonate material to put into them will be the nymphs. If they are started in the summer, they can be stocked more intelli- gently by obtaining the eggs of desirable odonates and hatching them. 10. Dragonflies will not feed, nor mate, nor lay their eggs in captivity, but damsel- flies are more susceptible, and eggs could probably be obtained from any common species. If a female dragonfly be caught while ovipositing and held by one pair of wings, leaving 248 BULLETIN OF THE BUREAU OF FISHERIES. the other pair free, she will deposit her eggs freely in any convenient receptacle if the tip of her abdomen be dipped in water. Thousands of eggs can be obtained in this way in a short time and kept until they hatch, or they can be placed at once in the breeding pond. The eggs of Anax, Aischna, and the damselflies can be secured by watching the females while ovipositing and then transferring the leaves or stems containing the eggs to the breeding pond. 11. If there is any necessity for rearmg the nymphs before placing them in the breeding pond, they can be fed on Paramcecium obtained by making an infusion of ma- nure in water, or on ordinary tow, especially the small crustacea, which they will devour in large numbers. Warren carried dragonfly nymphs successfully through their entire life history by feeding them with mosquito larve and pupe. 12. Whenever a fishpond is drained, the nymphs in it should be saved; they make excellent food for fish in other ponds and can be fed to them or can be used to restock the drained pond when it is filled again. 13. Dragonfly eggs hatch in 8 to 12 days; the nympa is short and thickset, the thorax and abdomen about as wide as the head, the legs long and slender, the antenne short and fairly stout, the eyes large, with black retinal spots surrounded by rings of colored pig- ment. ‘The mentum of the mask is much wider than long, with three mental sete on either side and a varying number of lateral and marginal setae. The lateral lobes have a terminal, movable hook, one raptorial seta, marginal sete on the outer margin, and usu- ally two small setee on the blade of the lobe. The respiratory tracheze are convoluted in the thorax and posterior abdomen and comparatively straight between the two, and are highly colored. 14. Damselfly eggs hatch in about three weeks; the nymph is long and slender, the thorax and abdomen considerably narrower than the head; the legs relatively short and slender; the antenne stout and long; the eyes small with few retinal spots, but each sur- rounded by colored pigment. The mentum is somewhat wider than long, with a single mental seta on either side and one or twolateral and marginalsete. The lateral lobes have a stout, terminal, movable hook and one raptorial seta. The respiratory trachee are highly colored and are convoluted in the thorax and anterior abdomen and are compara- tively straight posteriorly. The caudal gills are cylindrical, very long and slender, and taper regularly from the base to the tip. ANNOTATED LIST OF DRAGONFLIES AND DAMSELFLIES OBTAINED NEAR FAIRPORT, IOWA. THE GENUS GompHus.—The nymphs of this genus live in the mud or sand on the bottom of the Mississippi and its tributaries, and thus far none of them has been found in any of the fishponds. ‘There is no reason, however, why pond species like grasJinellus and submedianus should not be found there, as they probably will be in the future. They burrow into the mud and débris, leaving only the tip of the abdomen exposed for respiration, and lie in wait for their prey. They are both rapacious and omnivorous and will eat anything and everything smali enough to be caught and held by their power- ful jaws. They may be recognized by their thick and hairy, four-jointed antennz, which are usually inclined inward toward each other, by a flat labium simply folded beneath the chin, with strong grasping arms like mandibles and not extending up over the face in a DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 249 mask, and by the absence of dorsal spines along the midline of the abdomen. Their legs are stout and adapted for burrowing, the two front pairs directed forward and the posterior pair directed backward, and all three pairs armed at the tips with strong burrow- ing claws. Their color is similar to that of the débris in which they live, and the furry hairs covering their bodies and legs quickly gather a coating of mud which still further obscuresthem. When ready to transform they crawl up on some board surface like a float- ing barge, the side of a boat, the bark of a tree, or more commonly a flat mud surface close to the water. In suitable locations the mud will be found thickly incrusted with their nymph skins, and it is not uncommon to find two or more skins one on the top of another, those coming out last crawling up on the others. The imagoes scatter quickly as soon as they are able to fly and often entirely disap- pear from the vicinity where they leave their nymph skins. In general, the females retire inland, while the males remain along the water front. Their habits are very different from those of the Libellulidz, the imagoes usually alighting flat on the ground or close to it, or on the surface of a log, and squatting or flattening the body down until the wings almost touch the ground, all ready to spring upon their prey. ‘They almost never alight on twigs or grass stems, like the Libellulide. From this flattened position they dart out over the water, skimming close to the surface and going toward the center of the river or stream, and then return again to the bank. Their motion is more similar to the ordinary aeroplane than that of most Libellulids, and they often hover over one spot for some time. They pluck much of their prey off the surface of the water, and some species actually dive beneath the surface, entirely disap- pearing under the water. Although two of the species, externus and vastus, are common about the ponds, they have never yet been observed hunting their prey over the pondsurface. They do catch numbers of teneral dragonflies and damselflies, however, in the vegetation around the margins of the ponds. The imagoes are so nearly alike in color pattern, wing venation, and even in habits, that it is very difficult for anyone but an expert to distinguish the various species, but the nymphs afford much plainer distinctive characters. The pairing of the sexes occurs toward evening, and if one goes out just before sunset there is a good chance of obtaining the two sexes together. The male does not accompany the female while ovipositing, and the sexes never fly about together after the manner of some of the Libellulide. GOMPHUS PLAGIATUS Selys. Gomphus plagiatus Selys, Bull. Acad. Belgique, 1854, vol. 21, p. 57. In the Proceedings of the Indiana Academy of Science for 1901, page 123, Williamson said: ‘‘The why, whence, and whither of imago Gomphi is a puzzle.’’ And to no species of the genus apparently does it apply any more forcibly than to the present. The mymph skins of this species were by far the most abundant of any found along the river bank, and over 1,000 were collected during the summer. But the only imagoes caught or seen were two tenerals captured as they were emerging. The imagoes must depart as soon as they can fly at all, and apparently they remain in the place to which they go. Consequently this species can hardly be said to have any part at all in the ecology of the fishponds although such large numbers of them are transformed within a few hundred feet of the ponds. Emergence takes place early ia the season, usually during the night, so that by the next morning the imago can fly fairly well. 250 BULLETIN OF THE BUREAU OF FISHERIES. GOMPHUS VASTUS Walsh. Gomphus vastus Walsh, Proc. Acad. Nat. Sci., Phila., 1862, p. 391. This species was third in abundance, judging by the number of skins obtained, and it takes an active part in the odonate life around the ponds. While the imagoes emerge along the river bank, and while many of them remain there, others migrate to the vicinity of the ponds. Most of these migrants are females, although there is a respectable sprin- kling of males. And yet none of them ever deposits eggs in the ponds; they all return to the river. Hence the part which they play is strictly confined to the imagoes, and consists wholly in the consumption of various insects and the teneral imagos of smaller dragonflies. GOMPHUS EXTERNUS Hagen. Gomphus externus Hagen, Monogr. Gomphide, 1857, p. 411. This species was second in abundance, as shown by the skins collected. The sexes seemed fairly well divided, and specimens could nearly always be seen along the cinder road to the north of the ponds or in the vegetation on the embankments. They are active and restless hunters and voracious eaters. One female was observed July 18 eating an Argia putrida imago and was so intent upon her meal that she allowed an approach to within 2 feet. She chewed and swallowed every scrap of the large damselfly except the wings and had no sooner finished than she caught another and ate it similarly. Such gormandizing must of necessity play an important part in the ecology of the ponds. Like vastus, this species returns to the river for ovipositing, and its nymphs are never found in the ponds. Like plagiatus, emergence takes place during the night, and many tenerals were found early in the morning at the season of transformation, which seems to last through June and July. GOMPHUS SUBMEDIANUS Williamson. Gomphus submedianus Williamson, Entomol. News, 1914, vol. 25, p. 54. This species was found only at Patterson Lake and Sunfish Lake, on the Illinois side of the Missis- sippi River, just above Fairport. The males were plentiful along the banks of the lakes, while the females were found in swampy places some distance back in the woods. Specimens were sent to E. B. Williamson, the founder of the species, and he very kindly confirmed their identification. The males usually fly close to the surface of the water and have the habit of hovering for a short time over one spot after the manner of some of the other Gomphids. They also frequently alight upon floating logs, bushes, or some water plant. While hovering, the seventh, eighth, and ninth abdominal segments have a decided reddish tinge when the sunlight strikes them just right. GOMPHUS AMNICOLA Walsh. Gomphus amnicola Walsh, Proc. Acad. Nat. Sci., Phila., 1862, p. 396. This species was especially abundant along the banks of the river just above the ponds in series B and was occasionally captured around the fishponds. While the exuvie collected give us the best idea of the actual number of imagoes of the various species, the apparent abundance does not always corre- spond. The present species was seen and captured as often as any other single species, but in the number of exuvie it was far behind most of the other Gomphids. Evidently these imagoes do not migrate after their emergence, but stay around in the immediate vicinity. This dragonfly frequents the thick grass and underbrush a little back from the water’s edge and can be captured with comparative ease. The nymphs and nymph skins were all obtained from the river, and none was found in any of the ponds. GOMPHUS NOTATUS Rambur. Gomphus notatus Rambur, Ins. Neur., 1842, p. 162. This dragonfly is a little larger than amnicola, but has similar habits; it stays out in the open rather more, but is occasionally found in the thick grass. Its favorite haunt is along the river’s bank, whence it makes long flights out over the water, returning again to nearly the same place. The nymphs frequent the shallower portions of the river, and none are ever found in the ponds. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 251 NASIZSSCHNA PENTACANTHA (Rambur). Aeschna pentacantha Rambur, Ins. Neur., 1842, p. 208. This species is said to have a wide geographical range, but not to occur anywhere in abundance. It was found quite plentifully around Patterson Lake and the slew leading up to it from the river. Its habits are like those of other Aischnines; it patrols the banks, flying back and forth over a limited area and frequently alighting and clinging to the underside of twigs and branches, its body hanging vertically and its wings drooping. In this position it is not very difficult to catch. No old nymphs or nymph skins could be found, although careful and continued search was made for them, but newly hatched nymphs were obtained in August, 1917. ANAX JUNIUS (Drury). Libellula junia Drury, Ilust. Exot. Entomol., vol. 1, 1773, p. 112. According to Kellicott, this species is the first to appear in the spring and almost the last to dis- appear in the fall. During June the imagos were not very plentiful around the ponds, but they increased greatly in actual numbers and still more in relative abundance as the season advanced, and by the last of August they were surpassed only by Libellula luctuosa. Thisisone of the most powerful fliers and almost never alights except for ovipositing. At such times the two sexes fly about together and, alighting upon some water plant at or near the surface of the water, the male assists the female as she inserts her eggs in the tissue of the plant stem. Both sexes often have a regular beat which they patrol back and forth for a long time; they also fly later at night than any other species, sometimes high in the air, catch- ing the numerous small insects which they find there. Nymphs were found in all the ponds, but especially in 4 and 9, where they were not much dis- turbed by the fish. This nymph is probably better known and more often figured than that of any other dragonfly. It shows a great variety in its color pattern at different ages, as well as the usual differences according to the nearness of the next molt. When very small, it is a uniform greyish green; as it grows larger it becomes banded transversely with black and white, while the mature nymphs are bright grass- green, with a beautiful and intricate color pattern of cinnamon brown. Two medium-sized nymphs were taken in pond 4 that were snow white throughout and so transparent that the dark breathing trachee around the posterior intestine showed through plainly. The nymphs are most abundant in waters filled with vegetation, and may be found even in small ditches and pools, and there are sometimes two broods in a year. They expel the water from their rectum with a noise like that made in ejecting saliva, and such spitting served to locate most of them in pond 4 when the water was drawn. When those that were left in this pond transformed, they seemed to find the screen across the outlet peculiarly attractive, and it was covered with bunches of skins. Two of these bunches are shown in Plate LXVIII, figure 1, the right-hand one containing six skins in a row, each fastened to the one in front of it. fESCNHA CONSTRICTA Say. “Eschna constricta Say, Jour. Acad. Nat. Sci., Phila., vol. 8, 1839, p. rr. Nymphs were found in all the ponds associated with those of Anax; they do not transform until mid- summer or later, and hence no imagos are seen until then. The imagos frequently enter houses or other buildings and may often be captured there. They wander afar in the fields and are seldom seen around the ponds, preferring some small brook among the hills. They feed on flies as well as mos- quitoes and often catch house flies and stable flies around our dwellings. MACROMIA TAENIOLATA Rambur. Macromia teeniolata Rambur, Ins. Neur., 1842, p. 139. This and the following species were only found in the slews along the Mississippi River. None have ever been seen around the fishponds, nor have any nymphs or nymph skins been found there. Patterson Lake is a favorite resort of the imagos, but careful search in its waters failed to reveal any nymphs. 252 BULLETIN OF THE BUREAU OF FISHERIES. MACROMIA ILLINOIENSIS Walsh. Macromia illinoiensis Walsh, Proc. Acad. Nat. Sci., Phila., 1862, p. 397. Like the preceding species this one is never found around the ponds, but may be seen frequently along the river bank and at sunset in the comfields flying back and forth between the rows. At night both species congregate in favorite places upon low bushes and hang by their legs from the under side of the branches like Nasizschna. EPICORDULIA PRINCEPS (Hagen). Epitheca princeps Hagen, Synop.-Neuropt. of N. A., 1861, p. 134. This species is easily recognized by its large size, by the brown blotches on its wings at the nodus and stigma, and by the fact that there are never any white areas connected with these blotches as in L. pulchella. The males have regular areas which they patrol incessantly hour after hour, hawking the varied insect life they may find. The nymphs are common in all the ponds; but the imagos scatter after emerging, and only a few are seen about the ponds at any one time. The nymph is large and sprawling and can not cling well to grass stems, preferring a broad surface like a board, a stump, or even the side of a bank. Most of those taken at the ponds were found on a hard mud bank beside the cinder road. The two sexes do not fasten together during oviposition, but the female drops her eggs alone into deep water. The imago emerges early in the moming and is one of those that consequently falls a prey to the birds, since it is helpless during the first forenoon. Some of the wings of this species were found with those of L. luctuosa already noted (p. 222). PANTALA FLAVESCENS (Fabricius). Libellula flavescens Frabricius, Entomol. System. Supple., 1798, p. 285. This species was found more plentifully upon the Illinois side of the river, but was occcsionally taken around the fishponds. Nymphs were found in ponds 4 and 8, and skins were obtained along the shores of ponds 1, 2, and 3. The imagos are rapid flyers and very difficult to capture while on the wing; they congregate in open places near the river bank, where they may be recognized by their reddish-yellow color and strong flight. Apparently they never become really numerous anywhere in the vicinity of the station, but are one of the rarer species. PANTALA HYMENZAA (Say). Libellula hymenzea Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 18. Similar to flavescens, but with a distinct fuscous spot at the base of the posterior wings; common along the river bank just above the ponds in series B. Like flavescens they are rapid flyers and difficult to capture while on the wing, but, unlike that species, they frequently alight upon the under side of a twig of some bush or tree like Macromia and Nasieschna, and can then be captured easily. Nymphs were taken in ponds 3 and 4, but no skins were found in any of the counts made. Neither imagos nor nymphs occur in sufficient numbers to affect the ecology of the ponds. TRAMEA LACERATA Hagen. Tramea lacerata Hagen, Synop. Neurop. N. A., 1861, p. 145. This species can be readily recognized even when flying by the large black blotches at the bases of the posterior wings. The male accompanies the female while ovipositing, and the two may fre- quently be seen flying tandem over the ponds. Early in the season, June and the first of July, the species is comparatively rare, but later they become more numerous and by the last of August they share with Anax the honors of first place. Nymphs were found in all the ponds, and skins were obtained in every count made, those in August being especially numerous. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 253 TRAMEA ONUSTA Hagen. Tramea onusta Hagen, Synop. Neurop. N. A., 1861, p. 144. A single male of this species was captured on pond 8D in July, 1917, and others were seen later about several of the ponds. It can be readily distinguished from lacerata by the reddish color of the blotches at the bases of the wings. Neither nymphs nor skins were secured. PERITHEMIS DOMITIA (Drury). Libellula domitia Drury, Illust. Exot. Entomol., vol. 2, 1773, p. 83. This is the smallest of the dragonflies and may be distinguished from the others here mentioned by its diminutive size and its amber-tinted wings. The imagos are as common about the ponds as elsewhere, but are not very abundant anywhere. It is a slow and clumsy flyer, approaching more nearly to tie damsels than to the other dragons. Its small size gives it very little confidence and it keeps well out of the way of other species, flying close to the surface of the water. The female is usually found in the fields some distance away from the ponds, and she deposits her eggs unattended by the male. The nymphs are found sparingly in all the ponds and were also obtained from Patterson Lake on the river; they are cleaner, as well as smaller, than most other species. The skins are always found close to the water’s edge, often over the water, apparently on the first suitable stem that the nymph met with. CELITHEMIS EPONINA (Drury). Libellula eponina Drury, Illust. Exot. Entomol., vol. 2, 1773, p. 86. These dragonflies can be recognized by their heavily spotted wings and by their habit of balancing upon the very tip of some convenient grass or weed stem. When disturbed they return again to the same spot, and this makes them easy to capture. Their flight is slow, and in the position and movement of the wings bears more resemblance to that of a butterfly than of other dragonflies. They are seen paired and flying tandem more often than other species, and in spite of their slow flight they are more in evidence on windy days. The nymphs are found in all the ponds, while the skins are found close to the water’s edge, like those of Perithemis. The small size and scarcity of both species gives them but little influence in the ecology of the ponds. CELITHEMIS ELISA (Hagen). Diplax elisa Hagen, Synop. Neurop. N. A., 1861, p. 182. This species was first seen around the fishponds in the summer of 1917. One or two were seen in 1918, but none was captured. A male and female were secured on July 1, 1919, near ponds 4 and 8 of series D. LEUCORRHINIA INTACTA (Hagen). Diplax intacta Hagen, Synop. Neurop. N. A., 1861, p. 179. This is another small species familiarly known as “Johnny Whiteface’’; it may be recognized by its diminutive size and its snow-white face combined with a dark body and clear wings. The two sexes do not fasten together during ovipositing, but spend much of their time perched separately on some convenient object near the water. They fly only short distances from one resting place to another, but forage continuously all summer long. Their nymphs are found in all the ponds and much resemble those of Celithemis and Perithemis, but are shorter and generally show a definite color pattern of dark brown on a greenish background. They are lively and clamber about on the submerged vegetation with considerable agility. SYMPETRUM RUBICUNDULUM (Say). Libellula rubicundula Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 26. This species appears early in the season and remains until after the frosts of October. At first the adults are very soft and seem to remain teneral a long time, but later they become firmer and by Sep- tember are as rigid as any of the smaller species. 110307°—21——17 254 BULLETIN OF THE BUREAU OF FISHERIES. The imagos are found in large numbers around the ponds, but stick to the vegetation and do not fly out over the water. The nymphs were found in all of the ponds; and nymphiskins were present in all the counts. SYMPETRUM CORRUPTUM (Hagen). Mesothemis corrupta Hagen, Synop. Neurop. N. A., 1861, p. 171. Appears early in the season and is common around the ponds; then diminishes gradually and by the middle of August entirely disappears. Nymphs were found in all the ponds, and nymph skins occurred in the first two counts. It is the largest species of the genus and the strongest flyer, going out, like other dragonflies, over the water, but never in numbers, and remaining but a short time. ERYTHEMIS SIMPLICICOLLIS (Say). Libellula simplicicollis Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 28. Second in abundance at the ponds, sticking close to the grass and weeds and never taking long flights. It does not perch on the sides of the grass stems like L. Iuctuosa, but prefers a horizontal blade of grass and settles down flat upon it. The females hunt almost exclusively in the grass and feed upon diptera, small butterflies and moths, and damselflies, especially teneral Enallagmas and Lestes (Pl. LXIX, fig. 2.) Williamson (1899, p. 326) has noted a peculiar habit of the males. Two of them hover over the surface of the pond close to the water, one a few inches above and in front of the other. The lower one then rises in a curve over the back of the upper one, which at the same time moves in a curve down- ward, backward, and then upward, so that the positions of the two are exactly reversed. ‘The two keep this up for several minutes and then separate; stich movements may be witnessed on any clear day by watching for it. The two sexes never fly about together, but the female oviposits alone, hovering close to the water and repeatedly dipping the tip of the abdomen beneath the surface. Both sexes alight on the floating alge and other water plants, unlike most dragon flies. The nymphs never crawl far from the water to transform, and many of the skins are found upon rush stems standing in the water. They showed a curious preference for Carex stricta and Homalo- cenchrus oryzoides, and but very few skins were found on other plants. Such a preference was probably due more to the position of the plants than to any other factor. The two sexes are shown in Plate LXIX, figure 1. PACHYDIPLAX LONGIPENNIS (Burmeister). Libellula longipennis Burmeister, Handb. Entomol., vol. 2, 1839, p. 850. Not very common around the ponds, although a few can be found there all through the season. The matured, pruinose males are more in evidence than the females; both sexes have the habit of droop- ing the wings and elevating the abdomen when they alight. Nymphs were more abundant in ponds 2 and 3, and nymph skins more numerous in July and August. LIBELLULA LUCTUOSA Burmeister. Libellula luctuosa Burmeister, Handb. Entomol., vol. 2. 1839, p. 861. This is by far the most common species at Fairport and can be recognized by the broad, black bands across the wings, with chalky white spots outside of them in the male. It is very energetic and active, but alights often upon the grass and sedges and sometimes remains at rest a long time. It does not hover after the manner of some species and does not hunt late at night, being rarely seen actively flying about after sunset. It roosts in the tall grass up in the fields, holding onto the grass stem well down out of sight, and sometimes in the vegetation alongside of the ponds. Its characteristic attitude is to grasp the stem with all six legs, the longer hind legs holding the body inclined at an angle of about 45° with the stem, as shown in Plate LXVIII, figure 2. It gets thoroughly wet with the dew during the night and does not start flying in the morning until the dew has dried off. The two sexes do not fly about together after the manner of Anax, Tramea, and Celithemis, but the female oviposits alone, dropping her eggs loosely in the water, and not inserting them in the tissue of any water plant. Buty. U. S. B. F., 1917-18. PLATE LXVIII. Fic, 1.—Skins of nymphs of Anax junius left on the screen of the outlet of pond No. 4. Fic. 2.—Male of Libelluda luctuosa in characteristic attitude on the stem of a plant. Buy. U.S. B. F., 1917-18. PLATE LXIX. Fic. 1.—Male (upper) and female (lower) of Erythemis simplicicollis. Fic. 2.— A favorite resort of Erythemis simplicicollis beside pond No. 2. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 255 The males frequently come hawking around the laboratory building toward night, alight on the screens, and fly up and down the sides of the building, catching house flies, mayflies, and midges. They also have a curious habit of congregating around straw stacks in the open fields, probably attracted by the insects that frequent the sunny side of the stacks. Over the ponds they do not keep to a definite beat or patrol, but wander about indiscriminately, the males frequently clashing with one another. So far as observed, the imagos do not eat other dragon- flies or any of the larger insects. Large numbers of them while teneral fall victims to English sparrows and red-winged blackbirds. The nymphs are found in all the ponds, but most plentifully in ponds 3, 4, and 7. These nymphs are always dirty and without a color pattern, but the dorsal hooks on the abdominal segments are always visible. Sometimes they crawl long distances from the water, but most of the skins were found in the fringe of Carex close to the ponds. A few live in the Mississippi River, and these climb the willow trees on the bank and leave their skins attached to the bark. LIBELLULA PULCHELLA Drury. Libellula pulchella Drury, Illust. Exot. Entomol., vol. 1, 1773, p. 115. Common around all the ponds, but seeming to prefer those nearest the railroad and the ditch along the railroad track. It goes much farther from the water than the preceding species and is often found along the country roads and in the farmyards, industriously hunting the insects which occur there. Nymphs were found in all the ponds, and skins were obtained in every count, but were most abundant the last of July; they are among the largest of the Libellulid nymphs and make excellent fish food. PLATHEMIS LYDIA (Drury). Libellula lydia Drury, Illust. Exot. Entomol., vol. 1, 1773, p. 112. Like pulchella, this species prefers the ponds and the ditch along the railroad track; the nymphs were abundant in the ditch, but rare in the ponds. The male is easily recognized by his white pruinose body and black wings; the female has spotted wings and might be mistaken for pulchella, but is consider- ably smaller, and the triangle of the front wings is entirely free from color. This species is a persistent hunter, and the males have regular beats which they patrol almost constantly. DAMSELFLIES.—The habits of the various damselflies in ovipositing and the habits and relations of the nymphs to the fish life in the ponds are so similar that a general statement will cover them all, with the exception of a few peculiarities, which may be noted under the separate species. When ovipositing, the male grasps the female by the prothorax and flies about with her. She does not dip her abdomen beneath the surface and wash off the eggs after the manner of some dragonflies, but alights on some convenient water plant, floating alge, pond-lily leaf, or rush stem, or upon a floating twig or piece of wood, and places her eggs in position beneath the water, the male retaining his hold and assist- ing her out after she has finished. Often the male holds his body erect in the air, and floating objects are sometimes covered with the females busily ovipositing, while the males stand up from the surface like small twigs or moss stems. In some genera like Lestes, Argia, and Enallagma the female descends into the water and often draws the male in with her. The females of Argia putrida sometimes descend g inches beneath the surface, the female clinging to some water plant, the male holding his body erect, with the wings spread. After placing her eggs, the female releases her hold and the two rise to the surface, their buoyancy lifting the male into the air until his wings are free. He immediately begins to fly and lifts the female out of the water, and the two then go to another place and repeat the process. 256 BULLETIN OF THE BUREAU OF FISHERIES. However they may be laid, the eggs hatch quickly, and the ponds are swarming during the summer time with nymphs of all sizes and kinds. These nymphs have long masks which fold back beneath the head and thorax, like those of the dragonfly nymphs. But in place of the rectal respiratory apparatus of the latter, they carry three external tracheal gills at the posterior end of the abdomen. These are flattened laterally and are usually about half the length of the abdomen, their size and shape furnishing one means of identifying the species. Their food is similar to that of the dragon nymphs, but contains a larger percentage of small animals, as would be expected (table, p. 201). One such nymph was seen by Williamson (1899, p. 234) clinging to a dead catfish and evidently feeding on its flesh. ARGIA APICALIS (Say). Agrion apicalis Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 40. The male imagos of this species never become pruinose like putrida; one was secured from pond s, in which the blue of segment 8 was W-shaped and restricted to the base of the segment, like that in translata. The imagos were observed by Needham (1903, p. 242) at Galesburg, IIl., “feeding voraciously on adult Chironomids.’’ The mymphs were not found at all in the ponds, but were fairly common in the river. They do not travel far when ready for transformation, but the skins are always found within a few inches of the water’s edge. ARGIA MCSTA PUTRIDA (Hagen). Agrion putridum Hagen, Synop. Neurop. N. A., 1861, p. 96. The largest species of the genus; does not breed commonly in the ponds, but is very plentiful along the river. The imagos eat large numbers of mayflies, and when the latter are emerging almost every Argia, male and female, may be found munching one. The males quickly become pruinose, fading intoa uniform bluish gray, but the colors are usually restored on immersion in alcohol. The nymphs are numerous in the river, but only one or two were found in the ponds, and but few imagos were seen around the ponds. When ready for transformation the nymphs often go long dis- tances from the water and even climb rough-barked trees. Ten skins were taken from the trunk of a large willow tree 60 feet from the water, and with them were found half a dozen skins of Libellula luctuosa. AGRION (CALOPTERYX) MACULATUM (Beauvois). Agrion maculata Beauvois, Ins. Afr. Amer., 1805, p. 85. This beautiful damselfly is restricted to shady running water and is found only along a small brook one-fourth of a mile above the station. It sticks close to its haunts, although a male was seen one day fluttering along the shores of the ponds. Such visits, however, are only accidental, and the species does not enter into the life of the ponds to any appreciable degree. Tue GENUS ENALLAGMA.—Enallagma and Ischnura females, after inserting 8 or 10 eggs into the tissue of some plant, have a habit of stopping and straightening out the abdomen and stretching it, much as one stretches his fingers after prolonged writing. Evidently it requires considerable effort to thrust the ovipositor into the plant tissue, and since the abdomen is curved during the process it relieves the strain to straighten and stretch it. Two Enallagma females were observed on July 26 depositing their eggs. During the process each came in contact with a partially drowned damselfly floating in the water and tossed about by the waves, which they seized, pulled out of the water and ate. Here in the fishponds the Enallagma females seem to prefer the leaves of the crex grass as tissue in which to deposit their eggs. When the leaves break and fall over into the water, the part distal to the break dies and becomes apparently of just the right consistency to suit these damselflies, and nearly every such leaf contains eggs. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 257 A pair of Enallagma civile was observed upon a bullrush stem in pond 1D July 20, 1917. The female backed down the stem into the water for the purpose of laying her eggs. When the water reached the male and he became half submerged, he released his hold and perched on the stem above the water. But the female continued backing down the stem until she was at least 6 inches beneath the surface. Here she remained for ten and a half minutes actively ovipositing. Then a small sunfish, Lepomis euryorus, caught sight of her and snapped her up instantly. The female Enallagma often gets stranded on the surface of the water with her wings wet and unable to fly. When he catches sight of her in such a predicament, a male will fasten to her and try to pull her out. Such a resctte was witnessed in pond 2D; four different males fastened to this female, but the adhesion of the water was too strong for them. They could merely tow her along on the surface, each in turn giving way when he became exhausted. But together they pulled her far enough to reach some floating alge, onto which she crawled. Such chivalry deserved a far better reward than it received; a small cricket frog seized and swallowed her while she was drying her wings. ENALLAGMA ANTENNATUM (Say). Agrion antennata Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 39. This species is rare about the ponds, and only a few specimens were secured; elsewhere it is often found in large numbers and becomes the dominant species, as noted by Williamson (1899, p. 275). ENALLAGMA CALVERTI Morse. Enallagma calverti Morse, Psyche, 1895, p. 208. Only a single pair was secured from pond 4, the male of which could be recognized by the excellent figures given by Williamson (1900, pl. 1). ENALLAGMA CIVILE (Hagen). Agrion civile Hagen, Synop. Neurop. N. A., 1861, p. 88. This is one of the two most common species of the genus about the ponds, and its nymphs are found in every pond. In 1915 this species and Jschnura verticalis constituted the bulk of the damsel fauna of the ponds, but in 1916 there were fully as many of the species hageni as.of civile. Williamson stated (1899, p. 270) that old individuals of civile often have the wings milky or gray and the pterostigma bluish or pruinose, and this was noted in several specimens collected in September, 1915. Both the imagos and the nymphs take an active part in the life of the ponds, serving as food for fish and dragonflies. ENALLAGMA EBRIUM (Hagen). Agrion ebrium Hagen, Synop. Neurop. N. A., 1861, p. 89. Moderately abundant around the ponds and found in company with other species of the genus, which it very much resembles in habits and appearance. The nymphs were more abundant than the imagos and were found especially in ponds 1, 2, 3, and 4. They are just the right size to furnish good food for young fishes. ENALLAGMA GEMINATUM Kellicott. Enallagma geminatum Kellicott, Etom. News, vol. 6, 1895, p. 239. This is the smallest and most slender of the genus that frequents the ponds, but is also the most active, flying about restlessly over the water, often a long distance from the shore. It has the habit of sticking close to the surface of the water and alighting only on floating alge, which renders it difficult to capture. ENALLAGMA HAGENI (Walsh). Agrion hageni Walsh, Proc. Entomol. Soc., Phila., vol. 2, 1863, p. 234. This species, with civile and Ischnura verticalis, makes up 9o per cent of the damselfly life in and around the ponds. They are found everywhere in the vegetation near the ponds and often wander long 258 BULLETIN OF THE BUREAU OF FISHERIES. distances into the fields and woods. They are quiet and remain well concealed, so that often when none can be seen a sweep of the net through the vegetation will reveal them. ENALLAGMA SIGNATUM (Hagen). Agrion signatum Hagen, Snyop. Neurop. N. A., 1861, p. 84. This is the only orange-colored species of the genus found about the ponds, and this makes it con- spicuous while flying, as wellas atrest. The males and the two sexes when paired frequent the lily pads and similar water vegetation, sometimes long distances from shore. It is nearly as active and restless as geminatum and, like the latter, flies close to the water, making it difficult to catch. The species is common along the slews of the river as well as around the ponds and probably plays an important part in the life of those localities. HETARINA AMERICANA (Fabricius). Agrion americana Fabricius, Ent. Syst. Suppl., 1798, p. 287. This species frequents the neighborhood of flowing water, and hence is never found around the quiet ponds, but only on the river where the current runs swiftly. It is not very common anywhere near Fairport. ISCHNURA VERTICALIS (Say). Agrion verticalis Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 37. This is the most common damselfly around the ponds, appearing first in the spring and disappearing last in the fall. It frequents the thick grass and other vegetation to quite a distance from the margins of the ponds. The females are dimorphic; that is, of two different colors, and the black and the orange are about equally numerous. Both sexes are weak fliers and can be caught easily in the hands. The nymphs are abundant in all the ponds, and the stem of nearly every water plant projecting above the surface is covered with their skins. The species can be raised with little trouble if suitable aquaria are provided with the stems of rushes or similar water plants projecting above the surface. As the tables show, the nymphs are eaten not only by the fish but also by the larger dragonfly nymphs. There are probably a number of overlapping broods every season. LESTES EURINUS Say. Lestes eurinus Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 36. This is a large and stout species and has been found thus far only at the eastern end of the ponds; its nymphs have been taken in ponds 4 and 8, but not elsewhere. The imagos are associated with those of rectangularis and unguiculatus, and late in the afternoon the three species can sometimes be obtained in one sweep of the net. In spite of its size this species is a weak flier, but is an omnivorous eater. Two females were captured with partly eaten, brown moths in their mandibles; others were found eating butterflies, small caterpillars, and even teneral Enallagmas and Ischnuras. LESTES RECTANGULARIS Say. Lestes rectangularis Say, Jour. Acad. Nat. Sci., Phila., 1839, p. 34. This is a very long and slender species and may be recognized by these characters. It is the most common species around the ponds and was taken also along the river and at both Sunfish Lake and Patterson Lake on the Illinois side. It is rather more of a woodland species, but is found as well in the open, especially where there is rank vegetation to furnish shelter. It is not as omnivorous as the pre- ceding species, but feeds largely on gnats and midges. LESTES UNGUICULATUS Hagen. Lestes unguiculatus Hagen, Synop. Neurop. N. A., 1861, p. 70. The smallest of the Lestes species and found along only the eastern end of pond 4, where it is fairly common. No nymph that could be identified with this species was found in any of the ponds, although it is probable that the imagos referred to deposit their eggs in the ponds. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 259 LESTES UNCATUS Kirby. Lestes uncatus Kirby, Synonymic Cat., 1890, p. 160. A single male of this species was captured in one of the runs that stretch back onto the prairie August 10, 1917. The species can not be at all common in this vicinity, since this is the only specimen secured during four years of collecting. ISCHNURA (NEHALENNIA) POSITA (Hagen). Agrion positum Hagen, Synop. Neurop. N. A., 1861, p. 77. Needham (1903, p. 260) places this species in the genus Ischnura “chiefly because of the form of the abdontinal appendages in the male and the small round postocular spots.’’ Like verticalis, it appears early in the spring and continues until late in the fall, and its nymphs are associated with those of verticalis in the ponds. In habits and in their relation to the fish life of the ponds the two species may be treated as one. NEHALENNIA IRENE (Hagen). Agrion irene Hagen, Synop. Neurop. N. A., 1861, p. 74. This tiny species is associated with Lestes, frequenting the grass and vegetation in damp places. Like Ischnura, it feeds upon gnats and midges. Neither the imagos nor the nymphs occur in sufficient numbers to affect the life of the ponds. AMPHIAGRION SAUCIUM (Burmeister). Agrion saucium Burmeister, Handb. Entomol., vol. II, 1839, p. 810. A male and two females of this species were captured around the ponds below the railroad track July 26, 1917; none had ever been seen near the ponds in series D. However, this species is sometimes found in great numbers, and it may be that once obtaining a start here they will increase sufficiently to rank alongside the Enallagma species. BIBLIOGRAPHY... ATKINS, CHARLES G. 1908. Foods for young salmonoid fishes. Bulletin, U. S. Bureau of Fisheries, Vol. XVIII for 1908 (1910), part 2, pp. 841-851. Washington. Recommended different artificial:foods as the result of experiments upon young fishes. BAKER, FRANK C. 1916. The relation of mollusks to fish in Oneida Lake. Technical Publication No. 4, N. Y. State College of Forestry, Syracuse University, vol. 16, No. 21, pp. 8-366. Odonate nymphs were found in the stomachs of bullheads, bluegills, redeyes, common sunfish, yellow perch, manitou darters, and grass pikes. BALFOUR-BROWNE, FRANK. 1909. The life history of the agrionid dragonfly. Proceedings, Zoological Society of London, Ppp. 253-285, pls. 33 and 34. London. Two species, Agrion pulchellum and Ischnura ele- gans, were followed through their entire development from the egg to the imago. Details were given of every stage marked by the successive molts, and the development of the morphological form was fully discussed. The food on which the nymphs were reared was given in detail. BEAN, TARLETON H. 1912. Report of the fish-culturist for 1911. Annual Report, Conservation Commission for 1911, I, pp. 159-215. Mentioned dragonflies as an important trout food, p. 203. BENECKE, Dr. BERTHOLD. 1886. The enemies of pond culture in central Europe. (Translation by Herman Jacobson of ‘‘ Die Feinde der Teichwirtschaft’’ from “Die Teichwirtschaft,’’ Berlin, 1855.) Bulletin U.S. Fish Commission, Vol. VI, No. 22, pp. 337-342. Washington. Brré, L. 1884. Larve of Libellulide destroy 50,000 young fishes. Rovartani Lapok i, pp. 251-253 and xxx. Budapest. (Quoted in The Entomologist’s Monthly Magazine, vol. 21, p. 234, 1885. London.) Given in full and discussed on page 205. BRANDT, A. 1869. Beitrage zur Entwicklungsgeschichte der Libelluliden und Hemipteren. Mémoires, Académie Impériale des Sciences, série 7, t. xiii, no. 1, pp. 1-33, pls. 1 and 2. St. Petersbourg. Reared a parasite, Polynema ovulorum, from the eggs of Agrion (Calopteryx) virgo, and reported that some times half the eggs were destroyed in this way. Ca.vert, P. P. 1893. Catalogue of the Odonata (dragonflies) of the vicinity of Philadelphia, with an introduc- tion to the study of this group of insects. Transactions, American Entomological Society Vol. XX, pp. 1524-272, pls. 2 and 3, 2 text figs. Philadelphia. An excellent account of the structure of the imago and nymph, the life history, the geographical distribution, and the relationship of the Odonata, with a good bibliography. 1901. On Gomphus fraternus, externus, and crassus (order Odonata). Entomological News and Proceedings, Entomological Section, Academy of Natural Sciences of Philadelphia, vol. 12, pp. 65-73, pl. 3. Philadelphia. Wholly systematic. 1906. Nomenclature of certain North American Odonata. Idem, vol. 17, pp. 30-31. Showed that L. basalis Say, 1839, must be regarded as a synonym of L. luctuosa Burmeister, 1839. Campion, HERBERT. 1914. Some dragonflies and their prey. Annals and Magazine of Natural History (8), vol. 13, Pp- 495-504. London. Thirty-six records of various kinds of prey eaten by odonates; no American species included. 260 DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 261 CuapoT-KarLen. 1894. Report on the fish-cultural operations of M. Durand at the school of agriculture at Beaune. (Translation by F. P. Fennell of parts of report published by the National Society of Agri- culture of France, June 19, 1889.) Bulletin, U. S. Fish Commission for 1894, Vol. XIV, Pp- 309-310. Washington. Observations on) the rearing of Cyclops, Daphnia, and Cypris for fish food. Cypris was found to prey upon the young fish. Corer, R. E. 191s. Water conservation, fisheries, and food supply. Popular Science Monthly, vol. 86, pp. 90-99. New York. Quotation on page 185. 1916. The Fairport fisheries biological station; its equipment, organization, and functions. Bulle- tin, U. S. Bureau of Fisheries for 1914 (1916), Vol. XXXIV, pp. 385-405. Washington. Quotation on page 185. Dycue, L. L. 1914. Ponds, pond fish, and pond-fish culture. State Department of Fish and Game, Kansas, Bulletin No. 1, 208 pp. Topeka. Included both natural and artificial ponds, with directions for building the latter; described the habits and habitat of the basses, crappies, sunfishes, and catfishes, and gave directions for the artificial culture of each; discussed the enemies of fish and the problem of artificial feeding. ELrop, M. J. 1898. Iowan Odonata. Entomological News and. Proceedings, Entomological Section, Academy of Natural Sciences of Philadelphia, vol. 9, pp. 7-10. Philadelphia. A simple list of the Iowan dragonfly fauna. Emsopy, Geo. C. 1915. Thefarm fishpond. Cornell Readings Courses, Vol. IV, No. 94, pp. 213-252,4pl. Included types of ponds, location, water supply, pond construction, suitable fishes for the ponds, pond stocking, and pond management. Fisuer, A. K. 1893. The hawks and owls of the United States in their relationship to agriculture. U.S. Depart- ment of Agriculture; Bulletin No. 3. Washington. Forses, S. A. 1883. The food of the smaller fresh-water fishes. Bulletin, Illinois State Laboratory of Natural History, Vol. I, No. 6, pp. 65-94. Urbana. 1888a. Studies of the food of fresh-water fishes.. Idem, Vol. II, Art. VII, pp. 433-473. Peoria. 1888b. On the food relations of fresh-water fishes; a summary and discussion. Idem, Vol. II, Art. VIII, pp. 475-538. A summary of the two preceding papers, with additional notes and data. Garuicx, THEODATUS. 1880. A treatise on the artificial propagation of fish. Second edition, m. d., 128 pp. Cleveland. (First edition published in 1857.) One of the older treatises, but containing excellent suggestions. Garman, PHILIP. 1917. The Zygoptera, or damselflies, of Illinois. Bulletin, Illinois State Laboratory of Natural History, Vol. XII, Art. IV, pp. 411-587, pls. 58-73. Urbana. Morphology of nymphs and adults, life history, habits, and classification of Illinois species, with a good bibliog- raphy, keys for nymphs and females as well as males, tables, and charts. Hanxinson, Tuos. L. 1908. A biological survey of Walnut Lake, Mich. Michigan State Biological Survey Report for 1907, pp. 156-288, 75 pls. Lansing. Odonate nymphs were given among the food of the bullhead, redeye, blue-spotted sunfish, common sunfish, large-eared sunfish, bluegill, and largemouth black bass. 1910. Ecological notes on the fishes of Walnut Lake, Mich. Transactions, American Fisheries Society for 1910 (1911), pp. 195-206. Washington. A summary of the preceding paper with additional notes and data. 262 BULLETIN OF THE BUREAU OF FISHERIES. Hiccins, HELEN T. 1901. The development and comparative structure of the gizzard in the Odonata Zygopteza. Proceedings, Academy of Natural Sciences of Philadelphia, vol. 53, 3d series, pp. 126-141, pls. 2-4. Philadelphia. A detailed account of the folds and armature of the gizzard in go species of damselflies, nearly all foreign. Jounson, Rosert §., and StaPLeTon, M. F. 1915. Fish ponds on farms. Appendix II, Report, U. S. Commissioner of Fisheries for 1915, pp- 1-28, 19 pl. Washington. Included water supply, pond construction, value of aquatic plants, fishes suitable for pond culture, natural and artificial fish foods, diseases and enemies of fish. KAMMERER, PAUL. 1908. Symbiose zwischen Edogontum undulatum and Wasserjungferlarven. Wiener-Festschriften, pp. 238-252. Wien. Claimed a real symbiosis between the alga and the nymph, each deriving benefit from the presence of the other. Ket.icotr, Davip §. 1899. The Odonata of Ohio. Ohio State Academy of Science, Special Papers, No. 2, 114 pp., figs. 1-39. Columbus. An annotated list of Ohio species. KENNEDY, CLARENCE H. 1915. Notes on the life history and ecology of the dragonflies (Odonata) of Washington and Oregon. Proceedings, U. S. National Museum, vol. 49, pp. 259-345, 201 text figs. Washington. 1917. Notes on the life history and ecology of the dragonflies (Odonata) of central California and Nevada. Idem, vol. 52, pp. 483-635, 400 text figs. These two papers contain fine accounts of Pacific coast dragonflies and damselflies, with full descriptions and excellent illustrations. LAMBORN, Rost. H. 1890. Dragonflies vs. mosquitoes. 8vo., 202 pp., 9 pl. New York City. Introduction only by Dr. Lamborn. The dipterous enemies of man, by Mrs. C. B. Aaron; Utility of dragondies as destroyers of mosquitoes, by Archibald C. Weeks; The destruction of the mosquito, by Wm. Beutenmuller; Dragonflies as mosquito hawks on the western plains, by Capt. C. N. B. Macauley; Can the mosquito be exterminated? by Henry C. McCook, with good bibliographies. Lucas, W. J. 1908. Foe of dragonfly nymphs. The Entomologist, vol. 41, p. 16. London. “Mr. O. A. Rowden, writing from Exeter (England) on Dec. 16, 1907, says that the water boatman (Notonecta glauca) attacks the nymphs of dragonflies.”’ Lyon, Mary B. 1915. The ecology of the dragonfly nymphs of Cascadilla Creek (Odonata). Entomological News and Proceedings, Entomological Section, Academy of Natural Sciences of Phila- delphia, vol. 26, pp. 1-15, 1 pl., 1 diagram, 3 tables. Philadelphia. Included a census of the diatoms, green and blue-green alge, protozoa, etc., which settle on the back of odonate nymphs, the food of nymphs, and the efficiency of shelter as protection against enemies. MarsHALL, Wm. S., and GiLBert, N. C. 1905. Notes on the food and parasites of some fresh-water fishes from the lakes at Madison, Wis. Appendix, Report, U. S. Commissioner of Fisheries for 1904 (1905), pp. 513-522. Wash- ington. Odonate nymphs were found in the stomachs of 9 yellow perch. MartTIN, RENE. 1886. Les Odonates du département de l’indre. Revue d’Entomologie, vol. 5, pp. 231-251. Caen. McAtrE, W. L., and Brat, F. E. L. 1912. Some common game, aquatic and rapacious birds in relation to man. U.S. Department of Agriculture, Farmers’ Bulletin No. 497, 30 pp., 14 text figs. Washington. Mentioned as eating dragonfly imagos the killdeer plover, the common tern, and the black tern, and as eating dragonfly nymphs, the homed grebe, and Franklin’s gull; 327 nymphs were taken from the stomach of a single gull. DRAGONFLIES AND DAMSELFLIES IN PONDFISH CULTURE. 263 Meenan, Ws. E. 1913. Fish culture in ponds and other inland waters. 8°. Sturgis and Walton Co., New York. A general treatise taking up all the various aspects of the subject. MuirTKowskI, RICHARD A. 1908. Review of the dragonflies of Wisconsin. Bulletin, Wisconsin Natural History Society, vol. 6, Nos. x and 2, pp. 57-123, pl. and map. Milwaukee. Gave briefly a history of the Odonata, the life history, the geographic distribution, the seasonal distribution, the food habits and economic value, and directions for collecting and preserving. 1910. Catalogue of the Odonata of North America. Bulletin, Public Museum of the city of Mil- waukee, Vol. I, Art. I, 207 pp. Milwaukee. Included synonyms and all references of taxonomic value, as well as many of ethological and morphological character. NEEDHAM, JAMEs G. 1898. Birds vs. dragonflies. Osprey, vol. 2, pp. 85-86. 1899. Directions for collecting and rearing dragonflies, stoneflies, and mayflies. Bulletin, U. S. National Museum, Pt. O, No. 39, 9 pp., 4 text figs. Washington. “The best way to rear nymphs is to let them rear themselves;’’ does not inciude directions for feeding the nymphs. NEEDHAM, JAmEs G., and BETTEN, CORNELIUS. 1gor. Aquatic insects in the Adirondacks. N.Y. State Museum, Bulletin 47, pp. 383-612, 36 pls. Albany. An annotated list of the dragonflies (Anisoptera) of New York, giving especial attention to the nymphs, with extensive keys for both nymphs and imagos. NEEDHAM, JAMES G., and Hart, Cuas. A. 1901. The dragonflies (Odonata) of Illinois, with descriptions of the immature stages. Part I. Petaluride, Hschnide, and Gomphide. Bulletin, Illinois State Laboratory of Natural History, Vol. VI, art.1,94pp.,1pl. Champaign. Included an excellent general treatise on the Odonata, giving the literature, the life history, the habitat of the nymphs, the food relations, and directions for collecting and rearing nyimphs. NeEEpuHAm, JaMEs G.; MacGrutivray, ALEX D.; JoHANnNnsEN, O. A.; and Davis, K. C. 1903. Aquatic insects in New York State. N. Y. State Museum, Bulletin 68, Entomology 18, PP. 197-517, 52 pls. Albany. Included 7 parts of which Needham was the author of part 2, Food of brook troutin Bone Pond, and part 3, Life histories of Odonata, suborder Zygoptera. In the former the nymph of £schna constricta was said to eat trout fry, but was also found in the stomachs of two trout. In the latter were keys for both nymphs and imagos. NEEpDuHAM, JAMEs G., and Lxoyp, J. T. 1916. The life of inland waters. 8°, 438 pp., 244 text figs. Ithaca. An excellent account of the nature and types of aquatic environment, aquatic organisms, and their adjustment to environment, aquatic societies, and inland-water culture. Pearse, A. S. 1915. On the food of the small shore fishes in the waters near Madison, Wis. Bulletin, Wisconsin Natural History Society, vol. 13, No. 1, pp. 7-22. Milwaukee. Pou.ron, Epwarp B. 1906. Predaceous insects and their prey. Transactions, Entomological Society of London, pp. 323-409. London. A collection of original records of the food of various predaceous insects, including dragonflies, but all foreign species. Rity, C. V. 1888. Larva of Anax junius destroys carp. Insect Life, vol. 1, p. 58. Washington. Reference quoted in full and discussed on page 204. Se.ys-LonccHampes, Ep. DE, and HacEn H. A. 1850. Revue des Odonates ou Libellules d’Europe. Paris. Ssinitzin, D. Th. 1907. Observations sur les métamorphoses des trématodes. Archives de zoologie expérimentale et générale, t. 7, pp. 21-37. Paris. Found stages of a frog-lung fluke free in the body cavity of both nymphs and imagos of the damselfly, Agrion (Calopteryx) virgo. 264 BULLETIN OF THE BUREAU OF FISHERIES. TILLyvarD, R. J. 1917. The biology of dragonflies (Odonata or Paraneuroptera). Campridge Zoological Series, 8°, 396 pp., 188 text figs., 4 pls. A general treatise, including morphology, histology, embry- ology, geographic distribution, classification, geologic record, and bionomics, with an exten- sive bibliography. : WALKER, E. M. 1915. Noteson the Odonata of the vicinity of Go Home Bay, Georgian Bay. Ontario. Supplement to 47th Annual Report, Department of Marine and Fisheries, Fisheries Branch, Ottawa. Sessional paper No. 39, b, pp. 53-04, pls. 5-9. Gave ecological, seasonal, and geographical distribution of the species at Go Home Bay. Watsu, B. D. 1862. List of the Pseudoneuroptera of Illinois contained in the cabinet of the writer, with the description of over 40 new species, and notes on their structural affinities. Proceedings, Academy of Natural Sciences of Philadelphia, 2d series, 1862, pp. 361-402. Philadelphia. WARREN, ALFRED. 1915. A study of the food habits of the Hawaiian dragonflies or pinau, with reference to their eco- nomic relation to other insects. College of Hawaii Publications, Bulletin No. 3, 45 pp., 5 pls., 2 charts, 3 tables. Practically confined to the nymphs and imagos of Anax junius and Paniala flavescens; the food tables and charts are especially comprehensive and instructive. Wuiuiamson, E. B. 1899. The dragonflies of Indiana. Twenty-fourth Annual Report, State Geologist, Indiana, for 1899 (1900), pp. 229-333, pls. 1-7. Contained a general account of the life history, food, enemies, habits, and distribution of the Odonata, with directions for collecting and pre- serving them, and keys for the determination of both nymphs and imagos. tgoo. Additions to the Indiana list of dragonflies, with a few notes. Proceedings, Indiana Acad- emy of Science for 1900 (1901), pp. 173-178, 1 pl. Indianapolis. tgo1. Additions to the Indiana list of dragonflies, with afew notes. No.II. Idem for rgo1 (1902). Pp. 119-126, 1pl. ‘These two papers add other Indiana species, with notes and corrections. 1912. The dragonfly Argia mesita and a new species (Odonata). Entomological News and Proceed- ings, Entomological Section, Academy of Natural Sciences of Philadelphia, vol. 23, pp. 196-203. Philadelphia. 1914. Gomphus pallidus and two new related species (Odonata). Idem, vol. 25, pp. 49-58, pls. 4and 5. One of the new species was G. submedianus, which is common around Patterson Lake near Fairport. Wuson, Cuas. B. 1909. Dragonflies of the Mississippi Valley collected during the pearl-mussel investigations on the Mississippi River, July and August, 1907. Proceedings, U. S. National Museum, vol. 36, pp. 653-671. Washington. 1912. Dragonflies of the Cumberland Valley in Kentucky and Tennessee. Idem, vol. 43, pp. 189-200. Washington. These two papers are annotated lists of species, with especial teference to geographical distribution. Page. Abundance of species. ......... PARI SN 189, 19% Merisigryllus: i cisicicicsacc sR. EE 224 Zéschna brevistyla . 213,22 REINO EICHe Mere etomecoca ce cunines epranican ane ceen 223 ROTISERICEA see an cin ties ocacictetacdcetenacs 190, 205, 213, 233) 25% CUMMPA RS acer masiektinceme re cee rarest reac rere sa 224 multicolor. .......... SECCRER COOUROL SO cGORAIC AGbAetip: 223 Agrion (Calopteryx) maculatum.... .................-- 256 Pep LOY SUNT beige eeiter a AS MEE IDAD IAC aaeaa EMG aaa 26° 211 259 Sotelo tsls (aia sseeeeseress 187,190, 204,205, 213, 233,251 233 220 224 Aspia mpicalis ais His dae diel aes Soke ole ae ntcfenncereiad 215,256 Argia meesta putrida. . . PIPER CR LET DY AME ROS fo Ceili s walce's reicisiacietm asics cette te 215 RAPUICOMIA EY as sicise aaieeiants nedeiseinlice pe oe 210 Birds as enemies of imagos 222 Birds as enemies of nymphs 21 Palontervamcn ccs nce sc sgasieaeng caitiae aes a aetincgie ny ata 256 Celithemis elisa............... eee a ET tes 253 ROPETIA e s c sien diet aie winlc ate eater tcl t ciserhtatnercte Ceratopogon. . MPIC OA NOCA ES acie-tcia.ni- nic sino oiorsipaaialsi tin sins atvpimeeine 190 Choice of species. a 232 Gorncealment...5.2.6c00scenses sin si ain vive aieioiet tea wreietatom tees 189 Gordulegaster mactilatus, .. 0.0 ...0050scccseccscevteeces 205 Cordulia...... Soghecdcoger coeds im irefe sia Omrefetstetels ee eristeeinete 233 peer chtcsteccsieanichcie comme erictcrec/omteristetelocetnle sais 194, 202, 204 CECE a Re Bp ec IG OGRE SCcIeR rca Rac eiterecieycitess 187,188 (Cypris: <.5.5.. Morateiatbiotetabnlata aieipictalaisioty adie wisielotieiatatareistate’nicts 204, 208 WS CTU ONS BASES ABS aaa r BSB Er egeeob Siete reee ns 207 Or eT Es pooscdege gapanmriec ones eheiaiaaleiete aeoseg 204 claims of injuries... 204 202 202 238 Ior Diving beetles as enemies of nymphs isis) » WAKO: Dromogomphus spoliatus......... + 913,215 MEASURE SRM EMSIQU EN as isle) Ase jalste s\siateleipteinia orctelafd x siaiwie ce oferaie wnat 225 Dytiscus larvae... MoByi. m Me me i at | i 6 fi ’ * i —) , 7 te ee of teas Phares Ts tome Rar erin ie inet U is iL i sy =o Se | a a a t- zas , 4 ashe 4 = ‘ - 5 , oar V aie at he ‘ es fd wer 1a &! CONTENTS. A Page, Une 30-8 FGI DS ca.cacd 9300 Gao 708 OPO INU OCCU DORIC ODE DO ODDO OU CHOOSES ad OOO OSLO DO ORCUDUAn SOS: 269 GRIST PN RLV ER COL LECTINS ats aya siatsl eee '= loll tepstel nines eter octets seess/= yeas /shuola'a\ ajelo/ sie evele'aro/aiaVrale!aietats’ dpa 271 Dre Leaiatter aca rit ott te LPOG oeinc main he : 7. ~~ - aaa) Buu. U.S. B. F., 1917-18. Pate LXX. Fic. 1. Fic. 2. BURROWING MAYFLIES OF OUR LARGER LAKES AND STREAMS. 2 By James G. NEEDHAM, Professor of Limnology, Cornell University. & INTRODUCTION. In the beds of all our larger lakes and streams there exists a vast animal popula- tion, dependent, directly or indirectly, upon the rich organic food substances that are bestowed by gravity upon the bottom. Many fishes wander about over the bottom for- aging. Many mollusks, heavily armored and slow, go pushing their way and leaving trails through the bottom sand and sediment. And many smaller animals burrow, some by digging their way like moles, as do the young of mayflies and of gomphine dragonflies; some by ‘“‘worming”’ their way through the soil, as do the larve of crane flies and many oligochetes. Among the burrowers none are more abundant or more important than the young of the mayflies. Indeed, there are hardly any aquatic organisms of greater economic value, for they are among the principal herbivores of the waters, and they are all choice food for fishes. How abundant they are in all our large lakes and streams is well attested by the vast hordes of adults that appear in the air at the times of their annual swarming. They issue from the water mainly at night. They fly away to the banks and settle upon the shore vegetation. They cover the sides of buildings. They fly heedlessly into the faces of pedestrians. They settle upon the stream-side willows until their accumulated weight bends, and often breaks, the boughs. In the streets of riparian cities they fly to lights at night and fall beneath them in heaps upon the ground. Their bodies, crushed under the wheels of cars, render the rails slippery, sometimes impeding traffic. They feed a host of carnivores, terrestrial, aerial, and aquatic; indeed, many birds and fishes gormandize rather shockingly during their swarming season. And when those that have escaped both foes and casualties have conveyed their eggs back into their native waters, their bodies fall at last upon the surface and drift about. After the surfeited fishes can eat no more, the mayflies are blown into windrows upon the shores; or they drift in long lines that trail at the edges of the current in streams; or they gather in great masses and welter in the eddies. Sometimes they stop the river steamers by clogging the machinery. No dweller by the shores needs to be convinced of their abundance. That they abound also far out from the shore is well attested by certain observa- tions made by Commissioner Smith on Lake Erie nine years ago and later communi- eated in a letter. Dr. Smith wrote that, while making a cruise on a lighthouse tender which was visiting gas buoys in all parts of the lake, ‘‘Many of these buoys, especially 269 270 BULLETIN OF THE BUREAU OF FISHERIES. those toward the middle of the lake, had enormous quantities of dead mayflies on their flat surfaces. In several cases the mayflies formed a solid cake 6 or 8 inches deep, the result of one season’s accumulation.” That they are also of prime importance as food for fishes has long been known. The anglers found it out first and used the soft bodies of the larger mayflies successfully for bait. Indeed, in view of the lack of acquaintance with our mayfly species existing at the present day, the many common names for them that were used by fishermen of old seems surprising. Such old-time books for anglers as Ronald’s Fly-Fisher’s Ento- mology (1877) abundantly attest this. Food studies have everywhere demonstrated how generally the nymphs of these big, burrowing mayflies are eaten by fishes. Forbes’s report (1888b) was one of the most extensive. He found that these larve constitute nearly one-tenth of all the food taken by the fishes that he studied. He says (pp. 484-485.) : From the order Neuroptera fishes draw a larger part of their food than from any other single group. In fact nearly a fifth of the entire amount of food consumed by all the adult fishes examined by me con- sisted of aquatic larve of this order, the greater part of them larve of dayflies (Ephemeridz), principally of the genus Hexagenia. These neuropterous larve were eaten especially by the miller’s thumb, the sheepshead, the white and striped bass, the common perch, 13 species of the darters, both the black bass, 7 of the sunfishes, the rock bass and the croppies, the pirate perch, the brook silversides, the stick- lebacks, the mud minnow, the top minnows, the gizzard shad, the toothed herring, 12 species each of the true minnow family and of the suckers and buffalo, 5 catfishes, the dogfish, and the shovelfish—7o species out of the 87 which I have studied. Among the above I found them the most important food of the white bass, the toothed herring, the shovelfish (51 per cent), and the croppies; while they made a fourth or more of the alimentary contents of the sheepshead (46 per cent), the darters, the pirate perch, the common sunfishes (Lepomis and Che- nobryttus), the rock bass, the little pickerel, and the common sucker (36 percent). * * * The larve of Hexagenia, one of the commonest of the “‘river flies,’’ was by far the most important insect of this group, this alone amounting to about half of all the Neuroptera eaten. They made nearly one-half of the food of the shovelfish, more than one-tenth that of the sunfishes, and the principal food resource of half-grown sheepshead; but were rarely taken by the sucker family, and made only 5 per cent of the food of the catfish group. Forbes’s studies were made on such material as happened to be at hand, without regard to times or seasons or conditions under which the fishes he studied had obtained their food. Subsequent studies of the food of the shovel-nosed sturgeon have made it necessary to qualify his statement as regards that fish. Wagner (1908, p. 28) says: The food of Polyodon consists, in Lake Pepin, entirely of plankton material, in largest part of ento- mostraca, but not unmixed with alge. There is one seeming, but only seeming, exception to this. Occasionally one finds the specimens of a morning’s catch largely gorged with larvae of Ephemerids. But in every such case it was found that ephemerid imagines appeared in vast numbers the same even- ing. It appears plain, therefore, that the larvae taken by Polyodon were captured on their journey to the surface of the water. Wagner (1908, p. 31) adds concerning the valuable rock sturgeon: ‘The food of this fish in Lake Pepin, in the summer at least, consists entirely of the larve of the Ephemerids;” but he does not specify what sort of mayflies. Pearse (1915) records that 7 of the 16 species of small shore fishes studied by him at Madison, Wis., had eaten mayflies, the percentages of this sort of food averaging as high as 58.3 per cent in the largemouth black bass and 40 per cent in the bluegill sunfish. In regularity of consumption and in amount consumed the young of mayflies were second only to midge larva. But Pearse also fails to specify the kinds of mayflies BURROWING MAYFLIES. 271 eaten, which omission, in view of the very great diversity in size, form, and habits of different mayflies, detracts much from the value of his report. Materials for the present paper have been accumulated in the course of recent work done by various representatives of the Bureau of Fisheries in the Mississippi River. The most important part consists of field notes and specimens collected by Emerson Stringham in the vicinity of the big dam at Keokuk, Iowa. ‘This includes material for the life history of a species of Pentagenia, a genus whose immature stages have not hitherto been made known. An extensive collection of aquatic-insect larvae made by Dr. A. D. Howard in the course of a mussel-bed examination in Andalusia Chute, just above Fairport, Iowa, contains a considerable series of mayfly nymphs, mostly belonging to the group here treated. H. E. Schradieck collected adult mayflies on the grounds of the biological station at Fairport, Iowa, during the summer of 1916 and sent them tome. These adults are of the same species as are the immature stages from the mussel beds; thus they furnish additional evidence as to the mayfly population of the river. A. F. Shira, director of the station, sent me also the mayfly material from nine stomachs of the river herring (Pomolobus chrysochloris) collected by various persons. To this I have added data of my own, gathered during residence upon the shores of Lakes Mich- igan and Ontario and elsewhere. MISSISSIPPI RIVER COLLECTIONS. Since the pioneer work of Benjamin D. Walsh (1862, 1863, 1864) at Rock Island, Tll., the mayflies of the Mississippi have received little attention. Garman (1890) published a few observations on the habits of the adult and on the food® of the nymph of Hexagenia bilineata in the backwaters of the Mississippi bottom lands. But it has remained for the collections above mentioned, made by Emerson String- ham, H. E. Schradieck, and Dr. A. D. Howard, to add material data. Mr. Stringham’s collections, made by the waterside and along the Keokuk Dam, include adults, subi- magos, and nymphs taken at transformation. Mr. Schradieck’s collections, made from the walls of the biological laboratory at Fairport, Iowa, consist of adults that have flown from the river and alighted on the building. Dr. Howard’s collections, made with a dredge from the bed of the river in Andalusia Chute just above Fairport, are all nymphs. These three collections supplement each other remarkably well and give a better picture of the mayfly life of the river than we have had hitherto. Mr. Stringham’s collections were made between the middle of June and the middle half of September, 1916 (save for a few adults of an undetermined species of Heptagenia taken on May 29 and a single nymph of Siplonurus taken on June 7, and neither again recorded). They relate mainly to two species: Hexagenia bilincata Say (Pl. LXXI) and Pentagenia quadripunctata Walsh (Pl. LX XIII, fig. 15). Mr. Stringham’s notes were accompanied by specimens adequate for determination of the species. His record of the occurrence of these two species about the big dam is as follows: HEXAGENIA BILINEATA. JuNE 7.—One grown nymph collected. JuLy 12.—Yesterday I observed that a large, grayish-brown mayfly with two caudal sete was very abundant about the dam. Still more abundant to-day. a“ The food consists of earth richly charged with dead organic matter and with unicellular plants and animals. Such protozoans as Euglena are quite commoninit. A large part of the contents of the digestive tube is sand, which seems to be taken incidentelly”’ (loc. cit., p. 180). 272 BULLETIN OF THE BUREAU OF FISHERIES. Juty 13.—The species collected yesterday is much more abundant to-day. They swarmed all over me while I wason the dam. Every shaded place was covered with them. Juty 14.—Took photographs [see Pl. LXX, fig. 1] at about 5.30 a. m. of the gauge house on the dam between gates 29 and 30, covered with mayflies of which I collected specimens yesterday. In the lake above the dam there is a mass of cast skins, irregularly distributed, though the adults are distrib- uted along the dam very generally. Have not seen this species mating. From Keokuk to Montrose, Iowa, along the railroad track, the air was full of mayflies, all apparently the same species. The same conditions reported from Fort Madison. JULY 15, 6.30 a. m.—Only a few mayflies on the dam to-day—comparatively few. JuLy 17.—Mr. Howland (assistant general manager, Mississippi River Power Co.) tells me that he saw a pile of mayflies in Nauvoo, IIl., this morning that was about 6 feet in diameter and 18 to 20 inches deep. When I crossed the dam to-night at about 6.45 p. m. they were quite thick and were coupling and mating while flying. Above the chutes of the lock there is an enormous floating mass consisting mainly of mixed ephemerid adults, cast skins, and duckweed. JULY 18, 5.30 a. m.—As abundant as ever on the dam to-day. A little later —They are more abundant than ever this morning. JuLy 22.—Still a scattering of these on the dam this morning. JuLy 26.—Still a couple of thousand on the dam. Many on train arriving from the north; so it is evident they are still coming out on the lake. JuLyY 28.—Crossed the dam early this morning and was surprised to find thousands, perhaps millions, of the mayflies in the air. I believe that these are nearly ascommonasever. The north end of the power house (see Bureau of Fisheries Doc. No. 805, Pl. III) is blackened with them. Workman tells me this lot has been here about two days. Night Lock Master Raber tells me the present lot started last night. Collected some in process of transformation and some apparently just ready to transform. JuLY 30.—Large mass of dead adults above the locks; smell offensive on the lee side. Very few on dam this evening (6.15 p. m.), less than a hundred, I think. AUGUST I AND 2.—A few, scattering, on the dam. AucusT 5.—None on the dam to-day. Aucust 8, rr a. m.—A scattering of large, brown ones on the dam this moming—quite a number, in fact, for the late forenoon. The shady side of the poles had a dozen or two each; some posts had more. AuGusT 10.—A good many, but no great masses like those of last month. Collected a few at Fort Madison to-day. AucusT 12.—Collected some in Burlington, Iowa. AucusT 14.—As many on the dam as on the roth; possibly more. AUGUST 15 AND 16.—Mayflies still on the dam. Aucust 18.—Only a few dozen about the dam this morning. Aucust 19.—A hundred or more mayflies on the dam. AucGusT 21.—Noticed only one live mayfly. AucusT 23.—Again abundant on dam. Many on the gauge house; some on power-house walls; the horizontal portions of the dam thick with them, so that at each step one or more are crushed. Here- tofore they have been mostly confined to upright surfaces. Not very active; few on the wing, doubt- less because of cold. Aucust 27.—Did not see any mayflies when crossing the dam yesterday or early this morning. AuGUST 29.—Have not seen any more living ones. SEPTEMBER 2.—Some of the trolley poles on the dam, of which there are about 40, had 200 or 300 of large, brown mayflies; nearly all of them had some; few compared with earlier flights. SEPTEMBER 15, 8 a. m.—Noticed about two dozen clinging to floor of dam. [Observations discontinued.] Thus Mr. Stringham’s records show that during the summer of 1916, at least, July was the month of the principal flights of this species; that emergence was in waves; that successive waves reached their height at about the 13th, 18th, and 23d of the month, with falling away in numbers on intervening dates; that subsequent smaller waves culminated on the roth and the 23d of August, separated by intervals of entire BURROWING MAYFLIES. 273 absence of adults; and that belated reappearances occurred on the 2d and rsth of September. It is not likely that the first great wave arose on the 12th of July with such suddenness as Mr. Stringham’s notes indicate; it is more likely that a few adults appeared earlier but were unnoticed. Mr. Stringham noted the difference in behavior of the adult mayflies accompanying changes of temperature, adults when it is warm being able to cling to vertical surfaces, when it is cool (as on Aug. 23, minimum temperature 58° F., mean 70° F., some 15° lower than during preceding waves of emergence) lying flat upon horizontal surfacesonly. That the adults adjust themselves, also, in relation to light (Krecker, 1915) and to wind is well known. Desiring to know whether waves of emergence are influenced by local meteorological conditions, I requested data from Fred Z. Gosewisch, in charge of the Keokuk station of the U. S$. Weather Bureau. He very courteously sent me rather full data sheets covering temperature, precipitation, sunshine, winds, etc., and I have studied these carefully in relation to the facts furnished by Mr. Stringham, but I have not been able to trace any relation between emergence and meteorological conditions other than that which goes with the progress of the season. The mean daily tempera- ture at Keokuk for the month of June is 72.5° F.; for July, 77° F.; for August, 74.6° F.; for September, 66.4° F. Emergence mainly occurs at the hottest part of the season, but belated transformations trail along into the comparatively cool weather of early autumn. PENTAGENIA QUADRIPUNCTATA.4 JUNE 25.—They are transforming among the timbers of the boom above the lock. At 7.30 p.m. saw one leave cast (skin), and I took the insect and the cast. Took another just before it left the cast, but it struggled nearly free in the formaldehyde.® Saw more of these this evening than I have here- tofore seen this year. JUNE 27.—These are becoming more common. On a window sill over the water outside the com- pressor room, eastward of the lock, within an area of 114 square meters, I counted 31 of them at 6.15 a. m. Other windows similarly covered. At 9.30 a. m., on the entrance door of the power house (area about 3 Square meters), there were 53 of them. They were much thicker on this black iron door than on the dirty white concrete all about. JUNE 29.—Less abundant. JuLy 1.—To-night large numbers about the power house; also about the dam. JuLy 4.—On door of power house and on adjoining walls. Jury 11.—[No notes, but there is a vial containing a single nymph.] AucusT 12.—In Burlington, Iowa, yesterday and to-day, large, yellow mayflies; not very abun- dant. AuGusT 23.—Scattering on dam and on power house. AucusT 31.—Collected a solitary live one at 7 p. m. SEPTEMBER 2.—A few present. SEPTEMBER 7.—One on Fisheries launch No. 27 (ran from Fairport to Fort Madison, Iowa). Did not see any live ones about Keokuk yesterday or to-day. These less continuous observations on Pentagenia seem to show that it appears commonly about a fortnight earlier in the season than Hexagenia, and in smaller swarms, and that it continues to appear in dwindling numbers through the season. On July 4, Mr. Stringham wrote: Entrance door to power house had about 200 mayflies on it this morning. The adjoining walls were likewise covered. There appeared to be seven or eight different species, but the most common @ Again from Mr. Stringham’s field notes. > This specimen served for certain determination of the nymph, and for the life history which appears on a subsequent Page. 274 BULLETIN OF THE BUREAU OF FISHERIES. was a small one having brownish-gray wings, light-brown body, reddish antenne, and only two caudal sete. Accompanying this is a vial of the same date containing specimens of Pentagenia quadripunctata, Potamanthus flaveola, and Chirotenetes siccus. The commonest form was the subimago of the last named. This note has been quoted in full to call the attention of collectors to the diversity of appearance that may be presented by the different forms of a single species. Males and females differ strikingly in size of eyes, in length of legs and of tails, and in size and color; and each sex when adult may differ strikingly from its own antecedent subimago stage—the first winged stage that is assumed on leaving the water. ‘These differences are shown side by side for Chirotenetes siccus in Plate LX XXII. A single entry from Mr. Stringham’s notes (August 6), the only one applying to Polymitarcys albus Say, is quoted subsequently under the account of that species, page 286. Henry E. Schradieck, while engaged in other work at the Fairport biological labora- tory, at my request collected from the outside of the walls of the building, as he had opportunity, samples of the winged mayflies that settled there and kindly sent them to me in alcohol. The building is some 500 feet from the river and much nearer to the fish- ponds of the station, and these collections include a mixture of forms from both of these sources. Doubtless, Betis, Callibeetis, and Cenis came from the ponds, or from slack- water shoals, rather than from the open river. The other river species in the order of their abundance were: Chirotenetes siccus, best flyer of them all, July; Hexagenia bilineaia, July to October; Polymitarcys albus, September 6 to 11; Pentagenia vittigera, August 25; and Heptagema sp., July. Dr. A. D. Howard’s collections of nymphs were made with a fine-meshed dredging net that was drawn on lines at 25-foot intervals from the bank to the 1oo-foot line and from there on at 100-foot intervals. It covered several square miles of stream bed, and, although made primarily as a mussel survey, it furnishes much more comprehensive data of the insect life of the river bed under flowing water than we have hitherto possessed. Dr. Howard’s data will be published elsewhere; but it may not be out of place here to mention some facts concerning the river mayflies as evidenced by his insect collections that were sent me for determination. In this collection of over 600 specimens, sent in vials under 102 entry numbers, more than 75 per cent was composed of the following eight species of insect larve, in the proportions indicated: Speci- Times Speci- Times mens. occurring. mens, occurring, Dragonflies: Mayflies; Gomphus plagiatus............. ryt 190 57 Hexagenia bilineata.................. 38 12 Gomphus externus 67 37 Chirotenetes Siccus. < .ci4 0000505 .c0s 0 29 15 Stoneflies: Polymitarcys albus.................. 13 Acroneuria ruralis@ 8r 28 Pentagenia quadripunctata.......... 5 I Acroneuria abnormis 2x 12 @ The rupinsulensis of Walsh is a synonym of this species. The only other specimens sent in any considerable numbers were caddisworms of the genus Hydropsyche, of which 49 specimens were sent under 15 separate numbers. BURROWING MAYFLIES. 275 The mayflies are the chief herbivores among the insects of the stream bed. That they are not more numerous there is doubtless due to the remarkable abundance of carnivorous dragonflies and stoneflies always associated with them. In this population both pursued and pursuers fall into two principal ecological groups, according as they burrow in the sand and gravel of the stream bed or live in the water above it. The mayflies of the genera Hexagenia and Pentagenia and the dragonflies of the genus Gomphus are all true burrowers, possessed of flattened and more or less shovel-like, digging, front feet (see Pls. LXXII and LXXIV), have the hind legs appressed to the body and adapted for pushing, and have the front of the head sloping forward and somewhat pointed. There are also many special adaptations to burrowing, among which none is more remarkable than the development in these mayflies upon the front of the mandibles of a pair of long, strong, upcurving, and pointed tusks that are driven forward into the soil and upon which the roof of the burrow is lifted, opening a subterranean passageway. When a burrowing mayfly nymph is thrown out upon the surface of the sand, it digs in again more quickly than a mole. A few thrusts of the tusks forward, a few tosses of its head upward, a few side sweeps with its broad front feet, and it disappears from view beneath the sand. The mayflies of the genera Polymitarcys and Chirotenetes and the stoneflies of the genus Acroneuria live above the stream bed and do not burrow. They prefer the shelter of stones or of timbers but occur occasionally in more open places. In Dr. Howard’s collections these two mayflies are more frequently associated with one another than with any other species, and neither is once taken in association with either of the burrowing mayflies named above. Polymitarcys and Chirotenetes have little more in common, however, than a habitat. They are very different in form and also in manner of life. Polymitarcys is a bottom sprawler, depressed, flat, hairy, protectively colored, and inactive. Chiro- tenetes is an agile swimmer and an artful dodger, with body compressed, smooth, and of beautiful stream-line form. It gathers its food from the passing current by means of plancton-retaining fringes of hairs that margin the fore legs (Clemens, 1917). Poly- mitarcys is a member of the subfamily Ephemerine and has for its nearest allies the burrowing mayflies above discussed. Its mandibles are tusked (PI. LX XVIII, fig. 41), but the tusks are not upcurving and are not used for burrowing; they are laid out flat upon the bottom as are also those of the allied Euthyplocia (PI. LX XIX, fig. 48). Chirotenetes is a member of the subfamily Betine, a group in which there are no bur- rowers. This is the only member of the group that appears abundantly in the collec- tions from the Mississippi River, and, beyond the figures of Plate LX XXII illustrating Chirotenetes siccus, the group receives no further treatment in this paper. In January A. F. Shira, director of the Fairport station, sent me the insect con- tents of nine stomachs of river herring that had been collected during the two preceding seasons and observed to contain mayflies, so far as determinable, as indicated in the following table: 276 BULLETIN OF THE BUREAU OF FISHERIES. Foop or NINE SPECIMENS OF RIVER HERRING (POMOLOBUS CHRYSOCHLORIS). Hexagenia Pentage- - No. Date. Locality. Collector. nia Adults Hepta- Miscella- nymphs. x He = Es genia. | neous. ymPAS-! subima- ges gos.2 1 | May 23,1916 2| June 4,1915 3 | June 11,1916 4) June 21,1915 5 | July 2, 1916 : 6| July 4,1915 |..... GAT. ee tadssas ae Stringham. veeteees| Many. Many. c 7 | July 14,1916 | Lake Cooper above depot..|.....do... I I 209 c By) Ree MOS PACTS, Dear -beeit oes eae AUOW SIE? S215 duc halen 3 Too c 9) |: Sept, 8; xox6|, Keokuk, above darn sci o22| ccc sis wicisr teeiein|pciislan eae] Seoetetere matte J 100 c) @ Approximate only; many badly disintegrated. b The fishes named below were not sent me, having been previously determined. ¢ Occurring in large numbers. @ A damselfly nymph of the genus Argia. € x Hiodon and 4 undetermined fishes. / 2 Hiodon, 4 Dorosoma, 10 undetermined fishes. 9 Of recognizable specimens all were females. h 20 caddisflies of the family Leptoceridz. Evidently, during the season of flight of Hexagenia, this fish gorges itself with adults. Earlier it eats the nymphs. The eggs found might about equally well be obtained from nymph or adult, since they are matured, so far as external aspect is con- cerned, during the nymphal period. SYSTEMATIC ACCOUNT OF THE GROUP. The burrowing mayflies and their allies comprising the subfamily Ephemerine include in North America half a dozen genera of rather large species. Among these are the largest of our mayflies, the ‘‘brown drakes”’ of the genus Hexagenia, which by reason of their enormous swarms are known to everyone; the ‘‘yellow drakes”’ of the genus Pentagenia; the beautiful ‘“‘mackerels” of the genus Ephemera, with ornate color patterns on both wings and body, and most graceful and lively nuptial flight; and several genera of smaller and less familiar mayflies. These will be characterized and illustrated and an account of their habits so far as known will be given in the following pages. The group of the burrowing mayflies may be distinguished from other groups, and the genera of the group may be distinguished from each other in both adult and larval stages as follows: KEY TO THE SUBFAMILIES OF EPHEMERIDA. A. Adults. 1. Basal fork of the cubital vein 4 strongly unilateral; cubital and first anal veins strongly divergent EY La Sea OR a OCR Ean Sonia Sune ane Berne fe RRS ANS neko s gaat Boe Romolo EPHEMERINA. Basal fork of the cubital vein symmetrical, or nearly so; cubital and first anal veins at base parallel OPsveby slightly: cuverpentts: a crem er ier ceo) sate ntfs len eet 6 Batina: and HEPTAGENINA. B. Nymphs. 1. Mandibles with a prominent, tusklike, external branch projecting forward from the mouth and ‘visible from abovert,. oo. ASEESa a aor ee SSE GOCE: opate[etiets «a ar- ARE soya EPHEMERINA. Mandibles not; “tusked 7-5 ecse ari erie i eee cre lotr ieteda inte sta crete +b BaHTIna and HEPTAGENINA. @ The terminology of the venation of the wings is illustrated and explained in Plate LXXXI, figure ss. > Not here treated. BURROWING MAYFLIES. 277 KEY TO THE NORTH AMERICAN GENERA OF THE SUBFAMILY EPHEMERINA. A. Adults. 1. The posterior fork of the median vein in the fore wing very deep, almost reaching the base of the wing; two long simple intercalary veins between the first and second anal veins...... CaMPSURUS. The posterior fork of the median vein not extending more than three-fourths the distance to the Hase of the ‘wing..S2a20. See tt ANA OT SG PAIE ENS & MEP ONE Be ACN SOIT SRA Boies oe 2. 2. Between the first and second anal veins is a bunch of three or four long, straight, intercalary veins conjoined basally before their attachment to principal veins; the second anal vein is nearly straight, andgatibrahcheditye3s..228 Stat. 2021S BERS, RELA. ARR ER AL DI BO POLYMITARCYS. Between the first and second anal veins are only shorter, sinuate, and sometimes forking intercalary veins, that are attached directly to the first anal vein; the second anal vein is sinuate and often 3. The posterior fork of the median vein extends two-thirds to three-fourths the distance to the base of the wing; vein Cu, not more strongly curved at the base than is the first anal vein. .EUTHYPLOCIA. The posterior fork of the median vein less deep, not longer than its stem; vein Cu, more strongly curved:at-its:tbase thanasthe fitst/analiveim isis. oniwael.ctaheraige. att an cond. deal. gees... 4- 4. The third anal vein not forked, but attached to the hind margin of the wing by a series of cross veins; forceps of the’ male’4-jointed sia d3 gett woth GALICIA SCO « «208s oe se cede centages cs 5- The third anal vein ends in a single fork and is not attached to the hind margin of the wing, though a few isolated intercalated veinlets lie between; male forceps 3-jointed........... POTAMANTHUS. 5. Tails 3 in male and female; fore wings with a definite and beautiful pattern of spots. . EPHEMERA. Tails 2 in the male; fore wings diffusely marked or plain........ 0.0.0... cee cece cece 6. 6. Tails 3 in the female; mature color predominantly yellowish ..................... PENTAGENIA. Tails 2 in the female; mature color predominantly brownish ....................... HEXAGENIA. B. Nymphs. ye vUnksiown" (tropical and! subtropical) Aor. Jeo, OT I SB, POR 2B s CampsuRUS. Mandibular tusks shorter than the head, only their tips visible from above........ PoTAMANTHUS. Mandibular tusks longer than the head and very comspicuous........... 00.0000 ee cece ee eeees 2 2. Front of head rounded; legs decreasing in length posteriorly; fore legs longest................... Ze Labrum wider than long; tusks hairy almost to tips ............. cc cee ee eee eee EUTHYPLOCIA. 3. Front of head produced forward and conspicuously lobed; legs increasing in length posteriorly.... 4. Labrum longer than wide; tusks hairy only atenlarged base...................0020. POLYMITARCYS. aa hrantalspronunencessemucitcttiar, Smeliice 5.51) oi. eis weston cone cee seins seleelcadees HEXAGENIA. rantal aramiisienice: Dindvat ats iD sia daida le oe a aati ae Wigan Heels ohm ae es byeiiansts S: 5- 5. Both mandibular tusk and frontal prominence denticulate extermally.............. PENTAGENIA. Mandibular tusk and frontal prominence smooth externally ..............-.200---00005 EPHEMERA. SINGLE DISTINCTIVE CHARACTERS OF OUR GENERA OF &PHEMERINA. So well marked are these genera that they may be recognized at a glance by the following characters: Adults. EPHEMERA alone has the fore wings ornamented with a pattern of transverse spots. HeEXAGENIA alone has a border of brown on the front of the fore wing and another on the outer margin of the hind wing. PENTAGENIA alone has the transverse row of four dots on the veins, as shown in Plate LX XIII, fig- ure 15, and a single conspicuous dorsal stripe laid lengthwise of the body. CampsuRus alone has the middle and hind legs aborted, also the greater depth of the posterior fork of the median vein, almost reaching the wing base. PoTAMANTHUs is the smallest (expanse of wings about three-fourths of an inch), and it alone has the wings wholly transparent (the two following white-winged genera have gray or purplish margined fore wings). 278 BULLETIN OF THE BUREAU OF FISHERIES. PoLtymiTarcys alone has three or four long, straight, intercalary veins between the first and second anal veins, joined together basally before their attachment to these veins. Eutuypiocia alone lacks all the preceding characters. Nymphs. EPHEMERA alone has a frontal prominence divided by a deep, round notch into two smooth spines. HeExAGEnIA alone has a rounded, shelflike, frontal prominence on the head. PENTAGENIA alone has numerous brown denticles on margins of frontal prominence, antennal basal folds, tusks, and front tibie. These three are the true burrowers, having upturning tusks, front feet flattened for digging, more or less cylindric bodies, and erect gills. CaMPSURUS nymph is unknown. PoTAMANTHUS alone has the tusks shorter than the head. Potymirarcys alone has long, smooth tusks as long as the head and hairy only at their dilated bases. EvuTHYPLociA alone has the enormous tusks, hairy almost to the tip and beset also externally with brownish prickles. These last three are the sprawlers, having tusks horizontally extended, elongate fore legs, and laterally extended gills. HEXAGENIA, the Brown Drakes. This genus includes the largest of our mayflies, measuring often an inch and a half in expanse of wings and nearly an inch in length of body, to which the long tails may add 2 inches of length; all this without counting the very long fore legs which are usually extended forward. The fore wings are marked with a brownish band along the front border, and there is usually a narrower border of brown around the outer margin of the hind wings. This color varies from a faint, brownish tinge in newly emerged individ- uals and in pale varieties to dark brown, almost black, in older ones or in other varieties; and when these bands are darker, then the cross veins of the middle area of both wings become bordered with brown. The body is brown above, yellowish beneath; there is a paler longitudinal middorsal stripe upon the thorax, dividing the brown into two broad stripes, as in the typical H. bilineata Say; and there are interrupted yellowish rings upon the abdomen, all of which pale markings tend to become obscured in the darker specimens. Hexagenia bilineaia is the name I apply to all the variants of the species that occupies the beds of our larger lakes and streams. The color differences appear to be only differ- ences of degree. Even the differences of the male genitalia—usually our ultimate criteria of species—are intergradient. Walsh (1863) thought there were two good species in the Mississippi River at Rock Island. He said (p. 199): Nothing is easier than to distinguish the living specimens of these insects [H. bilineata and H. limbata=variabilis Eaton] by the color of the eyes. In the former the upper half of the eyes is cinnamon brown, in the latter bright, greenish yellow; in both the lower half of the eyes is black. The dried specimens, especially those of the male, are very difficult to distinguish. * * * In the middle of July, when on the shallow area of the Mississippi known as “the slough’’ at Rock Island, H. bilineata appears in prodigious swarms, so that the bushes absolutely bend down with their weight. * * * I am sure that in the thousands of individuals, both male and female, which blackened the bushes there was not one with the upper surface of the eyes yellow or yellowish; the only variation I noticed from the normal color was that one male had the eyes a shade or two paler than the rest on their upper surface. I have not had the privilege of studying the Hexagenias of the Mississippi River alive, but I am unconvinced by this emphatic opinion and by the long table of other BURROWING MAYFLIES 279 color differences that he gives on the following page (1863, p. 200); for I fear Walsh did not take into account the color differences due to age, and, though he examined thousands of specimens at a time, these thousands may well have been all of practically the same age—all of one day’s brood. Several forms have been named upon the basis of slight and inconstant color differ- ences. Dr. Hagen (1890) thought to reduce the species to two in number because of differ- ences he found in the form of the penes of the male, whether gently curved and finger- form (bilineata) or hooked and pointed (/imbata Pictet = variabilis Eaton). I thought for a long time I could recognize males of these two species; and in my earlier papers I have treated variabilis as a distinct species, but a careful study of more material has shown intergradients and additional forms. Four of the more typical forms of male appendages are shown in Plate LXXXI, figures 61, 62, 63, and 64. In a general way it may be stated that the long, straightish penis goes with the lighter coloration of wings and body and with northward distribution; but there are exceptions to this also. A separation of species on such characters as these should not be made without a careful study of at least two things: 1. Changes of form due to age.—A casual examination of the subimago, when the penis of the adult is clearly outlined within that of the subimago, shows that there will be a considerable change of form at the final molting. Figure 65 of Plate LXXXI shows this condition. The inclosed penis, it may be seen, will be of the form shown in figure 62 of the same plate. 2. Changes due to functional activity.—There is a relatively immense sperm mass gathered at the outlet of the vasa deferentia of the newly issued adult male, whose presence there may have something to do with the prominences of the penes and whose discharge may allow for much retraction. A good many names have been applied to the different forms of this genus, but after a careful study of a good bit of material from many localities I am unable to ree- ognize more than two good and distinct species in the eastern United States—a lowland species from lakes and rivers, Hexagenia bilineata Say, and an upland bog-stream species, H. recurvata Morgan.* The former is, of all our mayflies, the most important, the most abundant, the most observed, the most characteristic of the river fauna. The adult male is figured on Plate LX XI. The female would be similar, larger in size, with shorter fore legs and much shorter tails. The nymph is figured on Plate LXXII. I have examined nymphs and exuvie from many lake and river localities and have been unable to find any constant differences between them to indicate more than a single species. The material for this species that I have studied is as follows: Mississippi River material, as stated in detail in preceding pages, including adults from Emerson Stringham collected about and above the Big Dam, on dates ranging from July 12 to September 15, showing several great swarms in July and several reappear- ances in dwindling numbers to the end of the season; adults collected by H. E. Schra- dieck on the outer walls of the biological laboratory at Fairport, lowa, July 12 (when @ With this species we are not here concerned. It is still insufficiently described, but its determination is made possible by the figures of the genitalia that were published by Miss Morgan in the Annals of the Entomological Society of America, volume 6, page 395, 1913. Figures 8 and 12 of Plate LX XII herewith show that its nymph is readily distinguished from the lowland form by the relatively greater length of the mandibular tusks and by the unbranched condition of the rudimentary first gill. This is an early season species that swarms in May and then disappears, none being seen after early summer. 280 BULLETIN OF THE BUREAU OF FISHERIES. most abundant) and 25, and August 16; nymphs collected by Dr. A. D. Howard from the stream bed of Andalusia Chute just above Fairport, showing 12 occurrences in 102 col- lections; fragments of hundreds of adults (mostly females) from the full stomachs of the river herring, Pomolobus chrysochloris, sent by Director Shira. With these were a few nymphs that were probably taken when on their way to the surface to transform; and many nymphs taken from the stomachs of the shovelfish, Polyodon spathula, at Lake Pepin, and sent me by Dr. George Wagner, of the University of Wisconsin. These were doubtless taken while on their way to the surface, for they were eaten just before the swarms of adults appeared and at no other times (Wagner, 1908). The particular speci- mens sent me bear date of June 21, 1904. Other material made up of that collected by Prof. J. H. Comstock at Peoria, IIl., on July 17, 1887, and at Kidders Ferry on Cayuga Lake, N. Y., on July 17, 1886; by Prof. T. L. Hankinson at Walnut Lake, Mich., in June and July, 1906; by Prof. C. Betten at Buffalo, N. Y., on the 18th, the 24th, and the 31st of July, 1906, and in Storm Lake, Iowa, in June, 1902; by Prof. C. C. Adams at Ann Arbor, Mich., in June, July, and August, 1904; by Prof. George D. Shafer at Lansing, Mich., on August 16, 1906; by E. B. Williamson at Howe, Ind., on September 14, 1916; by Prof. C. R. Crosby at Co- lumbia, Mo., on June 16, 1905; and by myself in Lake Michigan, at Lake Forest, IIl., Walnut Lake, Mich., and Cayuga Lake, Ithaca, N. Y., July and August of several years. Hapits.—This species, though found in a wide variety of situations, prefers shoal waters whose beds are covered with soft ooze, rich in organic materials. There the burrowing is easy and food is abundant. Figure 2 of Plate LXX shows a soft-mud bank left bare by the receding stream. A portion of the mud has fallen into an undercutting current, and in the portion that remains undisturbed a section is exposed, perforated in all directions by the many burrows of the Hexagenia nymphs. Miss Morgan (1913, p. 99) has thus described the burrowing of the upland species which she studied: The sloping banks were mined by Hexagenia nymphs, the open burrows showing only 2 or 3 inches apart. Most of the burrows were apparent by their round openings; from some, hairy caudal sete pro- truded at full length. When a nymph was pulled out it speedily began to burrow again, placing the fore legs together with the bladelike tarsi held vertically. It next pressed them forward and outward, at the same time wedging the head between them in the cavity thus made. This movement was followed by a sudden lurch of the body forward, accompanied by wriggling of the abdomen. During these mo- tions the second pair of legs was folded close up to the body, while the third pair was held outstretched, ready to brace against the mud. These motions, rapidly repeated, enabled the nymph to bury itself in a surprisingly shorttime. Some of thesoft ooze taken from where the burrows were most numerous was later examined in the laboratory and found to be packed with diatoms. Stomachs of two of the nymphs were found full of silt and diatoms, showing that the nymphs had found plentiful forage as they burrowed. The length of nymphal life is unknown, possibly two years, if one may judge by the half-grown nymphs one finds in midsummer. ‘Transformation occurs at the surface of the water and usually at night. The grown nymph swims up and floats. A rent appears in the skin of its back. The subimago suddenly emerges from this rent, its wings ex- panding full size almost instantly. It stands a moment on the surface and then rises and flies away to the shore. It settles on any convenient support, often alongside countless others of its kind, as figure 1 of Plate LXX testifies, and remains quiescent for about 24 hours, when it molts again and becomes fully adult. Probably on the even- ing of the day following its final molt (I do not know that this has been determined in BURROWING MAYFLIES. 281 any case) it flies forth over the waters with myriads of its kind, all together making a great company, filling the air and forming one of the great swarms that have been so often described and that are so well known on every stream side. Mating occurs in the air during flight, and almost at once thereafter the female seeks the surface of the water. She flies hither and yon, dipping the surface, and then falling flat upon it with wings outspread. Her eggs are liberated in the water, just when and how has not been actually observed; but I have seen many playing above the sur- face of the water, egg-laden, and I have picked many from the surface, all of which have been spent females with eggs all gone. When gravid females are injured, as by squeezing the thorax or snipping off the head, they at once extrude their eggs in two long, yellow packets of extraordinary size (similar to those of Polymitarcys, Pl. LXXVII, fig. 36). If these be placed in water, the masses crumble and the eggs tend apart in falling; thus they become disturbed on the bottom. The egg coat, which is slightly adhesive, quickly gathers a protective covering of silt and becomes well-nigh invisible. The eggs of a single female will usually number above 8,000. The body of the female mayfly has become hardly more than a scaffolding for carrying this mass of eggs. Her mouth parts are atrophied; her alimentary canal is an air reservoir; her muscles are nearly all muscles of flight; her chief appendages are outriggers for control of flight; and her body is filled with eggs from end to end, even up into the rear of the head. To produce this great mass of eggs and to get them fertilized and back safely into the water, is her great end in life. The habits of the young nymphs that hatch from these eggs have not been observed. Doubtless they have many enemies, such as the predacious burrowing gomphine drag- onfly nymphs. They have also parasites. I found a large nematode worm filling the body cavity of one Mississippi River specimen, and most of the nymphs have their gills thickly beset with the cysts of some parasite unknown to me. The grown nymph of Hexagenia bilineata may briefly be described as follows : Length, 28 mm.; tails, 12 mm. additional; antenne,6 mm. Body pale, becoming purplish brown on abdomen and on gills, bare and shining on top of thorax, hairy around all margins. Frontal promi- nence of head shelflike, elliptical in outline when viewed from above, marked with a median dark dot, its margin fringed with pale hairs. There isa densely hairy circular ridge or fold surrounding the bases of the antennz externally. The long, strong mandibular tusks are bare and shining brown in color at their extreme tips, but bear a marginal line of hairs externally, the fringe becoming longer and denser toward the base. Prothorax with its side margins widened by a fringe of long, horizontally extended hairs. Fore legs stout and twisted. Femur ovoid, with a small lobe beside the apical articulation, hairy in longi- tudinal patches, the brushes short on the sides and very long on the edges. Tibix greatly dilated apically and further widened by marginal fringes of hairs, with a single, large apical tooth and that close beside the base of the short cylindric tarsal joint. Claw very short and stout, not more than twice as long as wide. Middle and hind legs slender, similar, each with a pincherlike prolongation of the apex of the tibia inferiorly, beside the base of the tarsus, the prolongation bearing an obliquely placed comb of short, stiff hairs. All expanded margins of all the legs bear dense brushes of yellow hairs. The hind wings of Keokuk nymphs show a distinct outer border of brown. Abdomen purplish brown above, darker on the middle segments, on each of which is included a pair of oblique, pale marks that are divergent at the rear. Gill on abdominal segment 1 rudimentary, tuning-fork shaped; on 2 to 7 large, composed of nearly equal lanceolate, long-tapering divisions that are broadly margined with whitish respiratory filaments. Tails stout, tapering broadly, fringed each side with tawny hairs, the very slender tips bare. 282 BULLETIN OF THE BUREAU OF FISHERIES. PENTAGENIA, the Yellow Drakes. This is another genus of very large mayflies. These are almost as large as the brown drakes but have shorter legs and tails. The prevailing color is yellowish, but there is a wide dorsal band of obscure brownish laid upon the body its entire length; the wings are only tinged with yellowish along the front border. There is, I think, a single species of this genus in the Mississippi River, though Walsh described two. I call them all, therefore, by the name he first used, Pentagenia vittigera; for Pentagenia quadripunctaia, the form with four dots on the veins of each fore wing (figured on Pl. LXXIII herewith), appears to be only a variant. The con- spicuousness of these dots is exaggerated in the figure; in a series of specimens some will be found in which one can hardly tell whether the dots are present or absent. This genus is very insufficiently known. Nothing has been published hitherto con- cerning it except bare and incomplete descriptions of the aduit. Mr. Stringham’s brief notes, cited in the preceding pages, are the first that deal with habits; and his care in collecting subimagos, together with their cast nymphal skins, make it possible to iden- tify with certainty the immature stages. The nymph is described below and is figured on Plate LXXIV. My specimens of Pentagenia are all from the Mississippi River. (It is reported in Banks’s Catalogue elsewhere only from Kansas.) The dates of adults range from June 16 to September 7, with maximum occurrence in late June. Transformation was observed by Mr. Stringham at 7.30 p. m. on June 25. Mr. Schradieck’s imagos from Fairport, Iowa, bear date of August 16. Dr. A. D. Howard encountered the nymph in the hed of Andalusia Chute only once. There is a single nymph sent me besides, bearing the label H. McAdams, Keokuk. The nymph of Pentagenia is similar to that of Hexagenia but is more yellow and more hairy, and a glance at the denticulations of the front of the head is sufficient for certain identification. It may be described as follows: Length, 24 mm.; tails, 6 mm. additional; antenne, 4 mm. Color pale yellowish, deepening to brownish posteriorly on the dorsum, whitish below. The head in front, the legs exteriorly, and the lateral lobes of the abdomen are densely clothed with shining, golden hairs. Eyes and ocelli black. The diffuse brown of the thoracic dorsum becomes a broad definite longitudinal stripe on the abdomen, nearly as wide as the segments. On each segment a pair of obliquely placed pale dashes is included in the brown. Antennz wholly pale and naked. Head with a strongly chitinized, two-toothed, frontal promi- nence, and a ridge bearing two unequal denticles above the base of each antenna. Cheeks and outer edges of palpi densely clothed with brushes of yellow hair. Mandibular tusks very strong, bare and upcurving in their apical third, dilated basally and bearing an external carina that is edged with irregular, brown denticles. This portion of the tusk is hairy within, and at the base externally. Top of head and thorax smooth, shining, save for some short, yellowish pubescence about the wing foots. Legs short and heavy, hairy on front and rear margins, especially the front legs. Front femora flattened and quadrangular, with an expanded inferior lobe beside the knee joint that bears an immense brush of stiff, yellow hairs. Front tibia dilated apically and bearing a similar inferior lobe that is densely clothed with shorter, stouter bristles; bearing also an external denticulate-margined ridge of increasing prominence distally where it ends in a strong and conspicuous brown tooth. Just beyond this, at the apex of the tibia, another similar tooth lies against the base of the short cylindric tarsal BURROWING MAYFLIES. 283 segment. There is an angulate line of long, thin hairs extending across the tibia near its base on the inner side. These hairs are outspread, fanlike beneath the mouth, meeting from the two sides, and may serve as a sort of table to hold up. the food convenient to the jaws. The front tarsus is hardly longer than the tibia is wide; its claw, a third as long, is short and thick and curved and abruptly tapering. Middle and hind legs are more slender and less hairy; third tarsi are longer, with claws of equal length, more slender and tapering. The wings of the nymph show four round, black spots at the points where dots will appear in the wing of the adult (see Pl. LXXIII, fig. 15), and two additional larger pigment spots that do not re- appear in the adult—one on the posterior fork of the median vein and one on the rear end of the humeral cross vein. The abdomen is nearly cylindrical, nearly covered above by the bushy, purplish gills. The lateral margins of segments 3 to 7 just outside the gill bases are expanded with bluntly rounded, lateral lobes that are densely clothed externally with golden-yellow hairs. These lobes increase in size to the fourth segment and then diminish in size posteriorly. The two divisions of the gill are unequal, the posterior being reduced in length, especially on the anterior gill-bearing segments. ‘The tails are rather short and stout and thickly fringed with hairs. EPHEMERA, the Mackerels. This is a genus of beautiful mayflies, somewhat smaller in size than the two pre- ceding, with relatively longer fore legs and tails. The cross veins of the front and middle portions of both wings are bordered with brown, and on the middle of the fore wing the brown is confluent in a series of spots arranged in a beautiful pattern. (See Pl. LXXV.)_ By this pattern our adult mackerels may all be easily recognized. The mackerels are lacustrine rather than fluviatile in habitat, especially E. simulans. The shores of the Great Lakes swarm with this species during early July, and on the Finger Lakes of New York they are only a little less abundant. The poorly differentiated species, E. varia Eaton, seems to prefer the little lakes and ponds and the muddy pools of stream beds. I have already published a description and figures of the nymph of E. varia (1901, p. 429). Miss Morgan has also figured it (1913, Pl. XLIV, fig. 8) and has added the following interesting notes on the habits of the species at Ithaca (p. 100) : No Ephemera nymphs were found in lower Fall Creek up to this time, that cleaner portion being nearly devoid of mud. On the first of July, however, the water in Beebe Lake was allowed to run off, bringing into the Lower Creek large quantitiesof mud. Three days later the shores below the dam were again examined. ‘Tracks similar to those made by earthworms covered the bottom near the shore line. Nymphs were crawling over the surface and sete could be seen projecting from many burrows. From an area of about 10 square feet 30 nymphs were removed. From this it appears that this species seeks out the muddy pools in even so turbulent and rocky a stream as is Fall Creek. E. varia is but doubtfully distinct from E. simulans. The wings of the typical form are less suffused with brown; the tip of the penis is a little more squarely truncated, and the proportionate length of the segments composing the forceps is slightly different; but none of these differences is either very tangible or very constant. I present new figures of the adult of E. varia on Plate LX XV, of its nymph on Plate LX XVI, and of the male genitalia of the three nominal species of the northern United States (these two and the very distinct E. guttulata) on Plate LX XXI, figures 58-60. Concerning the habitat and habits of the beautiful, but rare, broader-winged species E. guitulata Pictet no information is available. I have not seen that species alive, and the only specimen I possess is a fine male that was collected for me. 110307°—21——_19 284 BULLETIN OF’ THE BUREAU OF FISHERIES. So, again, in this genus there is a single species commonly occurring in the larger lakes and streams; and to that species mainly the following remarks will apply. I have observed Ephemera. simulans during several seasons at each of two places on the Great Lakes—at Lake Forest, Ill., on the shore of Lake Michigan, and at “ After- glow,” my summer home near North Fair Haven, N. Y., on the shore of Lake Ontario. In both of these places the Jake bed is sandy and not muddy. In neither place have I seen Hexagenia at all. Ephemera is the dominant form and the sole representative of the group here discussed, and it is associated in its swarming with several mayflies of other subfamilies and with a number of exceedingly abundant caddisflies. The swarms of Ephemera simulans arise less suddenly than do those of Hexagenia and decline more steadily after a single maximum that is reached about the first week in July. For many days together the herbage of the shores and the trees upon the bluffs are thickly besprinkled with adults for more than a stone’s throw inland from the shore. Then, after sundown, when they rise to enter upon their nuptial flight, the air is dark- ened with the clouds of them that extend in an unbroken line along the margin of the water. At Lake Forest, with an on-shore breeze their cast skins accumulate beside the piers in great floating masses, acres in total area and several inches deep, each cubic inch of these masses representing scores of individuals. At “Afterglow” the little hop hornbeam trees that cling to the front of the bluff are constantly aflutter with the mackerels that can not find resting places without disturbing one another; and if one shake such a tree, a perfect cloud of them will rise in the air The smallest lake at which I have studied E. simulans is Walnut Lake in Michigan. It is there, where the numbers are not so great and where the evening swarms are less rough-and-tumble, that I have seen their mating flight at its best. I have described this (1908), p. 261) as follows: After sundown the beautiful mayfly, Ephemera simulans, appears in companies of males over the edge of the water. The flight of one of these companies is a most delightful performance to witness, it is so light and graceful, and appears, withal, so exhilarating. Yet it is all up and down in vertical lines. With upturned head each individual flies rapidly upward, mounting quickly to a height of 10 or 15 meters; then spreading its wings out horizontally it falls upon them, with long fore legs extended forward and longer tails extended backward full length, rudderlike, keeping it always head to wind. Thus it descends, floating on the air, yet not drifting, until at the lower level of the swarm (4 or 5 meters above the water), it lifts its head and rises rapidly again in flight. And the whole company flying and falling thus, weaving up and down in vertical lines, and passing and repassing each other, create a scene of great animation. My material in this genus comes from a good many sources, none of it, however, from the Mississippi River, though doubtless the genus will yet be found in many places in that stream. Typical specimens of E. simulans come from Dr. C. C. Adams, collected at Portage Lake, Washington County, Mich., on May 30 (these bear the earliest date of all); from Prof. J. H. Comstock, collected on Cayuga Lake on the 1st and 6th of July; from Dr. C. Betten, collected at Buffalo, N. Y., on Lake Erie on the 11th of July; and from Wal- nut Lake, Mich., Lake Forest, Ill., and North Fair Haven, N. Y., on numerous dates in the fore part of July, collected by myself. Typical specimens of E. varia Eaton are from Three Mile Island, Lake Winnepe- saukee, N. H., collected by J. H. Emerton on July 10, 1906; from Lansing, Mich., col- lected by George D. Shafer; from Gloversville, N. Y., collected by Dr. C. P. Alexander BURROWING MAYFLIES. 285 on July 26; and from many places in the Adirondacks and about Ithaca, collected by myself. Miss Morgan gives the date of the first appearance at Ithaca as June 14. Of E. guitulata the sole specimen I have seen comes from Sport Island in the Sacan- daga River, N. Y., and was collected by Dr. C. P. Alexander. I have reared the nymphs of both E. simulans and E. varia several times, and have been unable to find any differences between them except a slight, somewhat intangible and apparently not quite constant difference in the shape of the notch that divides the frontal prominence into two points. This notch is in E. simulans typically a complete half circle, while in E. varia it is somewhat more widely open and forms but a segment (somewhat less than half) of a larger circle. Aside from this, the following description will apply to either species: Length when grown, 18 mm.; tails, 8 mm. additional; antenne, 4.5 mm. Color yellowish; abdomen with a pair of longitudinal brown streaks laid on the yellow which they divide. Antenne slender, twice the length of the mandibular tusks, thinly hairy above on basal half and naked thereafter to slender flexuous tips. The frontal prominence ends in a sharp tooth at either side, the two separated by a rounded notch in front. Mandibular tusks long, slender, upcurved, brown in color, and nearly naked. Maxillary palpi long, slender, thinly hairy, yellow. Legs moderately stout and somewhat flattened and twisted, clothed with tawny yellowish hairs on all exposed edges. All femora oval, fore tibia moderately flattened, widened from base outward only to midway its length, then parallel sided with an obtuse, bristle-covered, apical angle, but with lo accessory apical tooth. ‘Tarsus more than twice as long as wide and more than half as long as the tibia. Hind tibia prolonged beside the tarsus into a forcepslike joint, which is nearly wanting on the more normal middle legs. Wing cover of the grown nymph shows a transverse series of spots, which are those of the adult more closely grouped. Gills mainly yellowish, purplish only along the main axis. On segment 1, a bifurcated rudiment; on segments 2 to 7, large, bushy, the two divisions of about equal size. Tails thinly margined with tawny hairs. POLYMITARCYS, the Trailers. These are mayflies of medium size, having broad, white wings. The fore wings bear a border of dull, purplish color along their entire front margin. The legs are rather short, except the fore legs of the male, which are very long. The tails are very long. There is a single known North American species, P. albus Say. An adult female with protruding egg masses is shown in Plate LXXVII. The nymph is shown in Plate LXXVIII. The best account of the species is that given by W. E. Howard (in Needham et al., 1905, pp. 60-62), who studied it at Ottawa, IIl., from which account we quote a portion: Nymphs of P. albus are abundant in both the illinois and Fox Rivers at Ottawa. These rivers flow at this place over bottoms of solid sandstone, with bars of loose sand accumulated in the eddies. The streams are swift in the main currents, and the nymphs of this species are to be found under flat stones at the edge of swift water when about ready to transform. It was from two such situations that most of my collections were made, from which I succeeded in breeding a single specimen. I have seen the subimagos emerge and arise from the surface of the water in great numbers, but always just far enough out from the shore, so that the nymph skins were immediately swept into the current, where they disap- peared before they could be procured. The difficulty in collecting the skins from the natural breeding places is further heightened by the emergence occurring during the evening twilight. All emerge from the nymph skin at the surface of the water and leave the skin afloat. This is a midsummer species in northern Illinois. My bred specimen is dated June 22. None of the imagos in my collections shows an earlier date than this, but I have nymphs which are evidently 286 BULLETIN OF THE BUREAU OF FISHERIES. near to transforming which were collected the first week in June. Imagos and subimagos of the col- lections are scattered all through July, but August 5 shows them most abundant. At about this date they were observed in swarms. By the end of August they are much less numerous, and I have no col- lections which are as late as September. The subimago stage lasts 24 hours, and when the final emergence takes place the subimago alights on some object near the edge of the stream, where it transforms in less than a minute. The skin of the subimago remains attached to the bases of the sete of the imago and in this manner is carried out over the stream by the flying insect, where it is finally released after some minutes. Miss Morgan found this species in Fall Creek at Ithaca on June 20, the earliest date on record. I have received specimens from points on the Susquehanna River and from Corning on the Chemung River in southern New York, those from the latter place bearing date of August 20. Say found it swarming on Lake of the Woods, United States, Canadian boundary. Mr. Stringham’s single entry concerning this species is as follows: August 6.—Small, white mayflies have been common for at least a week or two about the river, though I have never seen any of themonthedam. Collected afew this evening at 6.45. In the moming there are many dead ones, but I never saw a live one in the day. His specimens bear date of August 5 and 6, and those of Mr. Schradieck from Fair- port, Iowa, bear the later date. So far as scattered records at present indicate, this species comes on slowly and reaches its maximum of swarming in early August. The species is probably much more widespread than present records indicate. Usually it is not abundant; it is pale and inconspicuous in coloration; it is quiescent in habits; it is crepuscular in flight; it is rarely noticed. The nymph does not burrow, but lies flat upon the bottom, with its legs and tusks and tails outspread upon the sand. It is protectively colored (see Pl. LXXVIII, fig. 39) and very inconspicuous. As noted before, it is an associate of Chirotenetes, and, like that species, its front tarsi bear two lines of long sete within, the hairs of the two lines diverging and extending forward when the legs are outspread. Observations are lacking on this species, but it seems probable that it also uses these fringes as strainers to gather plankton and other food out of the passing current. The full grown nymph may be briefly characterized as follows: Length, 15 mm.; tails, 7 mm. additional; antenne, 4 mm. Body depressed, widest across the thin, flaring, lateral margins of the prothorax, smooth. Color brownish, with faint marmorate markings on thorax, and a line of pale elongate spots laid upon and crossing the sutures between the abdominal segments; gills pale. Head broadly rounded. Antenne pale, bare. Mandibular tusk much shorter than the antenne, swollen at the base beneath the head, where clothed externally with many prickles and a few hairs, the tips long and slender and bare, very gently incurved. Prothorax widest across the front, where the thin, lateral margin is most expanded, and ends in a sharp angle directed forward just outside the rear of the head. Legs long, especially the fore legs, pale, but ringed and banded with brown and hairy along the edges; femora moderately flattened, twice banded; fore tibia and tarsus elongate and double fringed internally with long hairs; the long, straight, apical tibial spur is closely applied beneath the basal fourth of the tarsus, and a similar shorter spur at the tip of the tarsus is extended beside the sharply decurved claw. Abdomen depressed and with gills widely extended laterally. The gill on segment x is an erect, simple, hairy rudiment, on 2 to 7 double, composed of flattened tapering filaments. The lateral margins of the segments are rounded in front and angulate at the rear beside the gill bases. Tails rather stout segments ringed with short, pale bristles. BURROWING MAYFLIES. 287 EUTHYPLOCIA, the Flounders. This is a large, white-winged species, similar in aspect to Polymitarcys, but the dark, front border of the fore wings is more diffuse, showing a tinge of sepia or even roseate warmth of color, and the wing tips are hooked and so strongly corrugated lengthwise that the venation of the tips is difficult of examination. This is a tropical American genus likely to be taken only on our southern border. The type species E. hecuba Hagen is from Vera Cruz. The unidentified nymph of which I present a figure in Plate LX XIX, was collected by E. B. Williamson in Gualan, Guatemala. It appears to differ specifically from the un- identified nymph that was figured by Eaton on plate 29 of his monograph (1883-1886) ; however, that description was drawn froma cast skin; this, froma good alcoholic specimen. There are no other data accompanying these specimens. This nymph is remarkable for its flatness and for the extraordinary length of its tusks and antenne. This species may be briefly described as follows: Euthyplocia sp. Length, 29 mm.; tail, 12 mm. additional; antenne,15 mm. Color grayish brown, including the gills; antennal legs and sete yellow. Head and body depressed, widest across the prothorax. Head short and thick, depressed, wider than long, bare and shinning above, hairy about the mouth. Antennz very long, slender, flexuous, and bare. Mandibular tusks very long, sickle-shaped. Stout to near the tip when suddenly narrowed to bare brown points, hairy on both inner and outer margins, the outer margin and dorsal surface beset as well with brownish prickles. Prothorax broadly depressed with flaring parallel side margins; anterior angles more broadly rounded and incurved to a low, obtuse tooth at rear of head each side. Mesothorax with a low, thin, lat- eral lobe each side above the base of the middle leg, at front of segment, and strongly tapering rearward. Legs strong, thickly fringed with hairs in all exposed lateral margins, femora flattened and marked above with scarlike, longitudinal, bare areas. Fore tibia longer than its femur and prolonged still further by along, straight spine that lies closely beneath the tarsus for more than half its length. Tarsus similarly prolonged in a spur beneath the short, tapering claw. Middle and hind legs smaller, with only short apical prolongation of tibie. Wing cases uniformly purplish brown. Abdomen long, depressed, slowly tapering posteriorly, bordered by the wide fringe of the extended gills. A pale middorsal line emerges conspicuously on the posterior segments. The lateral margins also are pale, and there are obscure, paired, pale dots in the brown of the sides. Gills on segment 1 erect, simple rudiments; on segments 2 to 7 double, long, flattened, and copiously fringed with filaments. Tails very long, flexuous, and nearly bare. POTAMANTHUS, the Spinners. This genus includes the smallest and daintiest of our Ephemerine. They are white, faintly tinged with yellow in one species, P. flaveola Walsh, and with green in the male of the other, P. diaphanus Needham. They have an expanse of wings of something less than an inch, with white tails of the same length. There are minute fuscous markings on the tips of the segments of the fore legs of the male and on the middle cross veins of the fore wings of the female of P. flaveola that are entirely wanting in P. diaphanus. Only these two American species are known. ; The former (Pl. LX XX) is the species occurring in the Mississippi River. I have specimens of it also from Lansing, Mich., and from Ithaca, N.Y. Miss Morgan (1913) has published an excellent figure of the nymph, copied herewith on Plate LX XX] as figure 56. She says that ‘“‘in Fall Creek Potamanthus crawls upon silt-covered stones and muddy bottoms.’’ Eaton (1883-1886) says of the European P. luteus: ‘‘The nymph harbors 288 BULLETIN OF THE BUREAU OF FISHERIES. under stones in gently flowing water at the borders of rapids.’’ The other published ob. servations on the habits of the group are those of Betten (in Needham, 1908a, p. 194), as follows : Returning on the boat from Buffalo I happened to look up, and saw a swarm about 2ofeet above the water. I was able to take a few, but most of them were out of reach from the upper deck. It was too dark for me to see the manner of their flight. I returned next evening for further observation, but a strong wind prevented. I found the cast skins, however, belonging to this species floating upon the water and drifting upon the shore. The eggs of females of P. diaphanus in alcohol hang in rounded, globular masses be- neath the tip of the abdomen. My material in this genus ali bears dates in the month of July: P. flaveola, July 1 and 12 at Keokuk (Schradieck’s Fairport specimens have only the month specified); and P. diaphanus, collected in the Niagara River near Buffalo, July 31, 1906. The full grown nymph (Pl. LXXXI, fig. 56) may be briefly described as follows: Length, 13 mm.; tails, 4 mm. additional; antenne, 1 mm., their tips much surpassing the tips of the mandibular tusks. Body elongate and depressed. Prothorax wider than the head, with broadly rounded, flaring, lat- eral margins. Fore legs longer than the others; fore tibia much longer than femur, beset with long hairs internally and bearing a stout, straight, apical spur about half as long as the tarsus. Middle legs shorter and more slender than are the hind legs. Abdomen regularly tapering posteriorly; gills rudimentary on the first segment, well developed and about equal in size on segments 2 to 7. ‘he two divisions of each deeply fimbriate. Tails densely hairy along the middle portion, but bare at tips. BIBLIOGRAPHY. CLEMENS, W. A. 1917. An ecological study of the mayfly Chirotenetes. University of Toronto Studies. Biological Series No. 17, pp. 1-43, 5 pls. Toronto. (Published also as thesis, Cornell University, I915-) COLLINSON, PETER. 1746. Some observations on a sort of Libella or Ephemeron. Philosophical Transactions, Royal Society of London, vol. 44, pp. 363-366, 1 fig. London. DesmaR_Est, C. 1883. Notes on swarms of Polymitarcys virgo at Aubigne, France. Bulletin de la société ento- mologique de France (6), T. 3, cvii pp. Paris. Eaton, A. E. 1876. Notes on the legless condition of Campsurus. Proceedings, Entomological Society of London, pp. 7-8. London. 1883-86. A revisional monograph of recent Ephemeride or mayflies. Transactions, Linnean Society, London. Second series. Zoology, vol 3, pp. 1-352, 65 pls. London. 1892. Ephemeride. Biologia Centrali-Americana, pp. 1-16, 1 pl. London. Fores, S. A. 1888a. Studies of the food of fresh-water fishes. Bulletin, Illinois State Laboratory of Natural History, Vol. II, Art. VII, pp. 433-473. Urbana. 18885. On the food relations of fresh-water fishes; a summary and discussion. Idem, Art. VIII, PP- 475-538. Garman, H. 1890. A preliminary report on the animals of the Mississippi bottoms near Quincy, Ill., in August, 1888. Part x. Bulletin, Illinois State Laboratory of Natural History, Vol. III, Art. IX, pp- 123-184. Order Ephemeride, pp. 179-181. Peoria. Gentry, T. G. 1873. A swarm of mayflieson the Susquehanna River. Proceedings, Academy of Natural Sciences of Philadelphia, vol. 25, 3d series, p. 350. Philadelphia. BURROWING MAYFLIES. 289 Hacen, H. A. . 1861. Synopsis of the Neuroptera of North America. With a list of the South American species. Smithsonian Miscellaneous Collections, Voi. IV, Art. I, xx+347 pp. Washington. 1874. Report on the Pseudoneuroptera and Neuroptera collected by Lieut. W. L. Carpenter in 1873 in Colorado. Annual Report, U. S. Geological and Geographical Survey of the Territories for 1873. Part III, pp. 571-606. Ephemerina, pp. 578-583. Washington. 1890. Unser gegenwartige Kentniss der Ephemeriden. Stettiner entomologische Zeitung, red. v. Dohrn. Steitin, T. II, pp. 11-13. Berlin. Heymons, R. 1896a. Grundziige der Entwickelung und der Kérperbaues von Odonaten und Ephemeriden. Anhang zu den Abhandlungen der K. preuss, Akademie der Wissenschaften zu Berlin, 66 pp.,2 pl. Berlin. 18965. Ueber die Lebenweise und Entwickelung von Ephemera vulagta L. Sitzungsberichte der Gesellschaft naturforschender Freunde zu Berlin, pp. 82-96. Berlin. Krecker, F. H. 1915. Phenomena of orientation exhibited by Ephemeride. Biological Bulletin, Marine Biolog- ical Laboratory, Vol. XXIX, No. 6, pp. 381-388, 2 figs. Woods Hole. Morcan, ANNA H. 1913. A contribution to the biology of Mayflies. Annals of the Entomological Society of America, vol. 6, pp. 371-413, 13 pls. Columbus, Ohio. NEEDHAM, J. G. 1908a. Report of the entomologic field station conducted at Old Forge, N. Y., in the summer of 1905. N. Y. State Museum Bulletin 124, pp. 156-263, 32 pls. Albany. 19085. Notes on the aquatic insects of Walnut Lake. Report, State Board of Geological Survey of Michigan for 1907, pp. 252-271, 1 pl. Lansing. NEEDHAM, J. G., and BETTEN, CORNELIUS. 1gor. Aquatic insectsin the Adirondacks. N.Y. State Museum Bulletin 47, pp. 383-612, 36 pls., 42 figs., 2 maps. Ephemerida, pp. 418-429. Albany. NEEDHAM, J. G., and Luoyp, J. T. 1916. ‘The life of inland waters. S8vo, 438 pp., 244 figs. Ithaca. NEEDHAM, J. G., MorTON, KENNETH J., and JOHANNSEN, O. A. 1905. Mayflies and midges of New York: third report on aquatic insects. N. Y. State Museum Bulletin 86, pp. 1-352, 37 pls. Albany. Pearse, A. S. 1915. On the food of the small shore fishes in the waters near Madison, Wis. Bulletin, Wisconsin Natural History Society, vol. 13, No. 1, pp. 7-22. Milwaukee. -RONALDS, A. 1877. Fly-Fisher’s entomology. 8voed. London. Say, Tuomas. 1824. Noteson Polymitarcys. In Keating’s Narrative of an expedition to the source of St. Peter’s River under the command of Stephen H. Long, Major U. S. T. E., vol. 2. Description in app. vol. 2., p. 305. Philadelphia. Wacn_er, GEORGE. 1908. Notes on the fish fauna of Lake Pepin. Transactions, Wisconsin Academy of Sciences, Arts, and Letters, Vol. XVI, part 1, No. 1, pp. 23-37. Madison. Watsu, B. D. 1862. List of the Pseudoneuroptera of Illinois contained in the cabinet of the writer, with the description of over 40 new species, and notes on their structural affinities. Proceedings, Academy of Natural Sciences of Philadelphia. 2d series, 1862, pp. 361-402. Philadel- phia. 1863. Notes and descriptions of about 20 new species of Pseudoneuroptera. Proceedings, Ento- mological Society of Philadelphia, vol 2, pp. 182-271. Philadelphia. 1864. On the pupa of the Ephemerous genus Betisca. Idem, vol. 3, pp. 200-206. 1880. The same, translated into French and annotated by E. Joly. Bulletin de la société d’études scientifiques d’angers (Maine-et-Loire), pp. 157-174, 1 pl. 290 BULLETIN OF THE BUREAU OF FISHERIES. Watton, Izaax. 1653. The compleat angler; or the contemplative man’s recreation, pp. 91, 93, 97, 115. London. Wiu1amson, Hucu. 1802. Onthe Ephemeron leukon, usually called the white fly of the Passaick River. Transactions, American Philosophical Society, vol. 5, No. VIII, pp. 71-73. Philadelphia. EXPLANATIONS OF PLATES. PLATE LXX. Hexagenia bilineata Say. Fig. 1. The mayflies on the gauge house on the big dam at Keokuk, Iowa, at 5.30 a. m. July 14, 1916. They were about equally numerous all along the dam, which is nearlya milein length. Photo by Emerson Stringham. Fig. 2. The burrows of the mayfly nymphs in a mud bank of the Okaw River near Sullivan, II. Burrows exposed by lowering of water in the stream and by undercutting of the current giving a vertical section. Photo by Prof. T. L. Hankinson. PLATE LXXI. Hexagenia bilineata Say; adult. Fig. 3. The adult male (segmentation of the tails omitted, as in all the similar figures following). Fig. The legs of one side, fore, middle, and hind from left to right, as in the following plates. Fig. 5. The end of the abdomen of the male from beneath, showing forceps, penes, and base of tails; middle tail rudimentary. > Pirate LXXII. Hexagenia nymphs. Fig. 6. The nymph of H. bilineata Say; dorsal view. The hind feet are turned forward for compari- son with the other feet, and the gills on the right are moved aside to show the markings on the abdomen. The 3-branched, rudimentary gills on the first abdominal segment are abnormal. (See fig. 13.) Fig. 7. Portion of head of same, more enlarged. Fig. 8. Comparable portion of head of H. recurvata Morgan. Figs. 9,10, 11. Fore, middle, and hind legs, respectively, of H. bilineata Say; the front leg adapted for shoveling, the hind leg for pushing. Fig. 12. The unbranched rudimentary first gill of H. recurvata Morgan. Fig. 13. The normal bifid rudimentary first gill of H. bilineata Say. Fig. 14. The functional second gill of same, showing its fringes of respiratory filaments. PLaTe LXXIII. Pentagenia vittigera Walsh; adult. Fig. 15. Adult male of the quadripunctata form. Fig. 16. Legs of one side. Fig. 17. End of abdomen of male from beneath. Fig. 18. End of abdomen of female from beneath. PLATE LXXIV. Pentagenia vittigera Walsh; nymph. Fig. 19. The nymph from above, with the gills on the right pushed aside to show the markings of the abdomen. Fig. 20. Head of same from above, more enlarged. Fig. 21. Mandible of same, showing serrated tusk. Figs. 22, 23, 24. Fore, middle, and hind legs, respectively. Fig. 25. A rudimentary gill of the first abdominal segment. Fig. 26. A functional gill of the second abdominal segment. Fig. 27. Fig. 28. Fig. 29. Fig. 30. Fig. 31. BURROWING MAYFLIES. 291 PLATE LXXV. Ephemera varia Eaton; adult. The adult male. The legs of one side of the same. The end of abdomen of the same from beneath. PLATE LXXVI. Ephemera varia Eaton; nymph. The nymph from above, gills on the right tured aside to show normal color pattern. Head of same, more enlarged. Figs. 32, 33, 34. First, middle, and hind legs, respectively. Fig. 35. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. Fig. 39° A gill from the second abdominal segment. PLATE LXXVII. Polymitarcys albus Say; adult. Adult female, hovering with egg packets extruding. The legs of one side, adult male. End of abdomen of male from below, showing strongly divaricate penis tips. PLaTE LXXVIII. Polymitarcys albus Say; nymph. The nymph from above. Head of the same, more enlarged. Mandible of the same. Maxilla of the same. Labium of the same. Fore leg of the same. A gill of the second abdominal segment. PLATE LXXIX. Euthyplocia sp.? nymph. The nymph from above. Head of the same, more enlarged. Mandible of the same, hairs of the tusk omitted. Maxilla of the same. Labium of the same. A gill of the second abdominal segment. PLATE LXXX. Potamanthus flaveola Walsh; adult. The adult female. The legs of one side of the adult male. The ventral aspect of the male genitalia, showing the form of forceps and penes. PLATE LXXXI. Miscellaneous (figs. 55, 56, and 57 by Dr. Anna H. Morgan). The venation of the wings of Ephemera. Vein designations are those used in accompanying key; the six principal veins are: C. costa, Sc. subcosta, R. radius, 4. media, Cu. cubitus, A. anal vein. Branches of veins are marked with small numerals, in order from the front. The posterior fork of the median vein is the one in which veins M, and M, unite. R, is the radial sector; 1 is the principal accessory vein. The third anal vein is the forked vein behind the second anal. 292 Fig. 56. ays . 58. SOs . 60. SGX. . 62. - 63. 64: . 65. . 66. OT. . 68. ig. 69. - 70. ete AGED ergs BULLETIN OF THE BUREAU OF FISHERIES. The nymph of Potamanthus flaveola Walsh. An adult “ Peg-leg,’’ Campsurus sp.? (from Brazil), the extremely long tips of the tails omitted. Note the stubs of the vestigial middle and hind legs. Forceps and penes of Ephemera guttulata Pictet. Forceps and penes of Ephemera simulans Walker. Forceps and penes of Ephemera varia Eaton. Forceps and penis of Hexagenia bilineata, form bilineata Say. Forceps and penis of Hexagenia bilineata, form falcata n. nom. Forceps and penis of Hexagenia bilineata, form variabilis Eaton (syn., limbata Pictet). Forceps and penis of Hexagenia bilineata, form munda Eaton. Forceps and penis of Hexagenia bilineata, subimago, showing within the outline of the adult, form falcata. PLATE LXXXII. Chirotenetes siccus Walsh. The female subimago. Sketch of tails of male subimago to show their relative length. The female imago. Sketch of tails of male imago to show their relative length. Legs of one side of male imago. End of abdomen of adult male from beneath. The nymph from the side. Note the plancton-gathering fringes of the fore legs. The tail fin of the nymph; outer taiis fringed only on the inner side. a XI. . . . 7 8. AT BULL U » B. F., Ig! I SEs seria Sr LED a Gy Lies “) LO Spey Bury. U.S: Bees roxr7—18: PLATE LXXII. ‘2 eS Uj figs Kee WE AY NEES IN Bury. U.S. B. F., 1917-18. PLATE TL, XXIII. = aS =, — PLATE LXXIV. Buy. U.S. B. F., 1917-18. 2S > WMH); MYPLI A, Buu. U.S. B. F., 1917-18. PLATE LXXV. 5 m Yyy A | ful Yj a Sy P \ AW SSS =< —— taal HI = SSS su sulll NANI WH a > YN peers =" 7 } Son SS -7, a eS iil HM ae CB iiss HM Ld Lt MOBREGUI DDI Caw esse nccaceccaccices eee anand SS\G\ S Buu. U.S. B. F., r917-18. x | | PLATE LXXVII | Ss 4s seges “i N SEGRE NG LEE OSHA we eet ae Nl, \ nt ma) H : Le 27 y/ 36 BEA AO | Buuy. U.S. B. F., 1917-18. Pirate LXXVIIL. BuEL. U. Seba atom yale: HY Wj Whi YY} Yi We MY PLATE LXXIX. 10 Sere IN \ WIRNSSSS \\ SS | = FC SS Oo Ziyi CL4 LL] SS ZL LL AW WA SoMhy\\ SY Buu We. Dak. LOL7—US. Buu. U.S. B. F., 1917-18. PLATE L XXX. ———— ia ‘ f ba sso \\ ae AES Ath SS S SJ am a NA ile) ea, Ca Se I Oar a eae Bo 54 Bu. U.S. B. F., 1917-18. PLATE LXXXI. Bury. U. S. B. F., 1917-18. IBN) IL OOTUL. HABITS OF YELLOW PERCH IN WISCONSIN LAKES Sa By A. S. Pearse and Henrietta Achtenberg 293 Pee rT i i i E] 4 liga at he pn mayne gt wer ra ae amen ocean oe - Ht t Pt hs ¢ 2i¥Ai W2KODIW MI HORI WOLLAY 10 ZTIGAH mm * Are oe seeot 2 A va eisdasidcA siteinsH ; \ F H : CONTENTS. & Page divhs caval t (sia torr kone BOR SDB SOM REO SECTOR CHB ACIOATIE Mote MISE iat aa 8 Ons ars SNPS eo ann Dino ene 207 OO Hert yous of aee aisvaeier latrecosivieve rele siviciotns erelatam andl ME Ae POORER RTT ae oe ert 300 Qualitative and quantitative food determinations........... 0.2.22. e ee eee ence eee e nes 304 Season a Uivart atom erst nyaatesciate eta ohstero 21 shotcig sta (oreo tee cia niste set Natasa: Sins ia in -asBin aya elie eliayatia/ stern Sr ebvs avn 304 Ouantityotitood constimed and rate Of digestion ssh: 90. 80: 70. 6a. ae S50 - 40. me fey Lee ° 4 Jo. oe met == Za. i i 4 40. =i Lake Mendota, 1916. - - - - - - - , 24-inch mesh net at o to 5 m.; -.-.-.-.-, 1-inch mesh at o to 5m.; . I-inch mesh at 5 to Iom.; -0-0-0-0, t-inch mesh at ro tors m.;...... ,t-inch mesh at 15 to 20m. HABITS OF YELLOW PERCH. 303 ‘ <= a ry ee May June duly Aug. Sept. ocr Wor Ode: * ——. OO ee eee kee ee es SRC BEIG BARTS 10 LORE RES I 15 18202930 $14 2307 10 WIAD BLAST IEE? @ > : So 4s* v GO. JS - Jo 2s 20 5 \ ie Hoon of 4a* " 1 nee Xx Fis hit r » ieee é Ls alt ! % cf j t j \ aoe Qa 304 BULLETIN OF THE BUREAU OF FISHERIES. QUALITATIVE AND QUANTITATIVE FOOD DETERMINATIONS. Perch were usually examined while fresh, but in a few instances they were preserved in 95 per cent alcohol before examination, as in the case of those collected in Oconomowoc Lake. In the laboratory the contents were stripped from the alimentary canal on a microscopic slide. A little water was added to the mass; it was then teased apart under a binocular microscope and after being well spread out was again examined with a compound microscope. The volume of all constituents of the food was estimated and recorded in percentages. As a rule, the larger constituents were counted, and in many instances the number of microscopic animals was also noted. It was not practicable to measure the volume of the food, because it was mixed with more or less mucous secretion, so that in the intestine it formed a cylindrical “string.” The total number of adult perch for which we made such volumetric percentage food estimates in the lakes studied was 1,147. Considering together those of various localities, habitats, and ages, the food, as a whole, was made up of 38.3 per cent insect larve, 21.4 per cent entomostracans, 9.5 per cent insect pup and adults, 6.1 per cent silt and débris, 5.5 per cent macroscopic Crustacea, 5.5 per cent plants, 4.5 per cent fish, 2.4 per cent molluscs, 1.4 per cent oligochzetes, + leeches, + arachnids. The maximum amount and number of the particular species of animals observed in the food have also been recorded, and a number of important examples are given in Table 10. No particular discussion of the different items in the food is necessary. Tables 6 to 9 show that entomostracans and insect larvae are most important; but there is also a good representation of other animals, plants, and mud from the lake bottom. SEASONAL VARIATION. Though the diet of perch is made up of the same general kinds of food throughout the year, there is considerable seasonal variation in all the important constituents, some of which are eaten only during certain months. The seasonal appearance of various items as constituents of perch food is represented graphically in figures 6 to 30, the curves showing the fluctuations in each throughout the year. From a study of these the annual food cycle may be outlined somewhat as follows: Perch at all seasons feed largely on or near the bottom. During the spring they come inshore, probably chiefly for breeding, and feed more or less among the aquatic vegetation. This is indicated by the rise in the percentages of plants, gastropods, Corethra larva, silt, and fine débris in the food at that time. In summer perch leave the deep water on account of its stagnation and feed on the bottom near the thermocline, as is indicated by the increase of chironomid larve, crayfishes, and midge pupe# in the diet. After the autumnal overturn the perch return to deep water and feed largely on Cladocera and Corethra larve. During the winter they remain in the depths of the lake, as shown by the preponderance of cladocerans, silt and débris, chironomid larve, and Sialis larve. Some foods (like Corethra and chironomid adults, crayfishes, Corethra and chironomid pupz, mites, and ostracods) were eaten only during the warmer months; some (copepods, oligochztes, alge) were eaten throughout the year in small quantities; other foods (Corethra larve, chironomid larve, Cladocera, silt and débris, Sialis larve, etc.) appeared at all seasons but showed rather striking maxima during certain months. The time at which a particular food was taken in greatest quantity often coincided with HABITS OF YELLOW PERCH. 305 By) 2 Fig. 6. Fish fig. 7. x id Fig, 6. Ls 2s Chironomid ches Pupae ‘i. . et 20 j : 15 4a 5 o 5 Fig, , a Corethra Pupae ss a / is ne 0 os _ . Fics. 6 To 9.—Percentage by volume of four constituents of perch food which increased markedly in the summer. -o-0-0-, Lake Mendota, rors; - - - - - - , Lake Mendota, 1916; ......, Lake Wingra, 1916-17; , average. 306 BULLETIN OF THE BUREAU OF FISHERIES. FLL (Al geld AEA ee, WB, n 20 us Fig./0 " Caddis Fy He Larvae neh ‘ \ 15 we | \ i] \ il k | \ a ROA! r \ ! \ © ° \ 10 rh | 1 ' \ j \ { \ Sa, ¢ ! f . ; ot ‘ ! O = at psec: Fig. l/. Corerthra 45 Adults ad Ne cA v 4) crmalgee ae iT) : Fig, 12, OP at 2 raytishes if Qo ——= /0 a ang Fig. 13. “ \ Lamelibranchs 5 Fics. 10 to 13.— Percentage by volume of four constituents of perch food which increased markedly during the summer. -0-0-0-, Lake Mendota, to1s; - - - - - - , Lake Mendota, 1916; ....... , Lake Wingra, 1916-17; , average. HABITS OF YELLOW PERCH. 307 Jn tie PER AR TK A, AA AQ SMOOWNY @ 50 is 40 ae Fig. 14. es 30 CorethraLarva ; Fig 10. Ephemerid Nymphs * Se o . Fi 5 /6. . Odonata Nymphs fo Ae Fig. ff ue Neale Leeches oS ADO K 5 \ on fis 15 : if a Fig. 18. 40 Fics. 14 to 18.—Percentage by volume of five constituents of perch food which increased markedly during the spring or autumn, -0-0-0-, Lake Mendota, ro1s;, -- ---- , Lake Mendota, 1916; .....-, Lake Wingra, 1916-17; , average. 308 ep J 40 y Sj eA 4 7 eA Se 40 20 45 8S =~ dW Ss ~~ S ¢ ~ Som eo Fics. 19 to 24.—Percentage by volume of six constituents of perch food. -o-0-0-, Lake Mendota, tors; 1916; FUISMO AN IMS Fe FS BULLETIN OF THE BUREAU OF FISHERIES. (ye) Fig. /9, Chironomid Adults O” We 7 Fig.€9. ES ‘\ Gastropoda ,° \ x a i oN fig. 22. Ostracods fig. a9, Amphipods Fig. 24. oe a , Lake Wingra, 1916-17; , average. --+---, Lake Mendota HABITS OF YELLOW PERCH. 309 Poms eda gait: inte heel: REF cia” iis ater Aiahar Oe? PES, do Fig. 2s. 20 Sialis Larva Fig. 26. / Cladocera ME Fig. 27. Fics. 25 to 27.—Percentage by volume of three constituents of perch food which increased in amount in the winter. -0-0-0-, Lake Mendota, 1915; ------ , Lake Mendota, 1916; ...--- , Lake Wingra, 1916-17; , average. 310 BULLETIN OF THE BUREAU OF FISHERIES. J F ™“ 4A y . Cao | af A S a NV oO yan fia. 28. - \ Oligochaetes Fig. 30. HABITS OF YELLOW PERCH. 311 its period of greatest abundance (adult midges, midge pupz), but in other cases there was no such correlation. In the autumn the number of cladocerans in Lake Mendota increased (Birge, 1897), and the quantity eaten by perch also increased, but during the vernal cladoceran increase the opposite was true. Although copepods rivaled cladocerans in abundance at certain seasons they were never eaten in large quantities, probably because they are active, small, and do not collect in swarms to any extent, as cladocerans do. : The following list gives the number of perch out of the 1,147 examined which ate each constituent of the food; the percentage of the total which each constituent formed when it was 1 per cent or more; and the occurrence of the constituents throughout the year. All figures in parentheses mean percentage of total food by volume; figures outside parentheses indicate the number of fish eating each food. INSECT LARVA—Continued. Diptera larvee—Continued. Fisu (4.5), all year: Fish eggs— Unidentified, 2, April. Cisco eggs, 1, November. Perch eggs, 4, April. Sucker eggs, 4, May. Fish remains— Unidentified remains (3.9), 85, March to November. Abramis chrysoleucas, 1, October. Eucalia inconstans, 1, February. Lepomis incisor, 6, January, February. Notropis heterodon, 1, February. INSECT LARVA (38.3), all year: Diptera larve (32.7)— Unidentified chironomid larve (5), 148, all year. Chironomus abbreviatus, 10, May to Sep- tember. . decorus (8.3), 288, all year. . fulviventris (1.7), 86, all year. . lobiferus, 62, all year. - modestus, 3, April. . tentans (2), 28, October to February. . viridicollis, 7, October, December. . viridis, 1, August. . sp. 82 Johannsen, 1, May. . sp. 83 Johannsen, 3, May. Corethra punctipennis (5.6), 329, all year. Cricotopus trifasciatus, 7, April, May. Orthocladius sp., 3, July, March. O. nivoriundus, 2, July. Palpomyia longipennis, 6, May, July. Probezzia glaber, 12, April to July. Probezzia pallida, 14, April to July. Procladius sp., 29, March. Protenthes choreus (3), 166, all year. P. culiciformis, 2, August. Stratiomyia sp., 1, June. Tanypus sp., 37, all year. Qa000 aq0aaon T. carneus, 7, April, May. T. decoloratus, 6, April, May. T. monilis, 20, July, November. Tanytarsus dives, 15, March to May. Tanytarsus sp., 2, March. Ephemerid nymphs (1)— Mayfly sp., 32, all year. Betisca sp., 22, March to May. Cznis diminuta, 9, May to September. Callibetis sp., 18, April to May. Ecdyurus maculipennis, 1, June. Ephemerid sp., 14, May, June. Heptagenia interpunctata, 1, June. Siphlurus sp., 1, May. Odonata nymphs (1.4)— Damselfly sp., 34, all year. Argia sp., 1, May. Enallagma antennatum, 19, May to July. E. hageni, 28, March to July. E. pollutum, 6, May. Ischnura verticalis, 9, March to June. Dragonfly sp., 7, November to July. Anax junius, 2, March, July. Libellula sp., 1, April. Nehalenna irene, 1, April. Sympetrum, 3, October. Trichoptera larve (1.5)— Agraylea multipunctata, 10, June, July. Hydroptila, 3, June, August. Leptocella uwarowii, 17, March to Sep- tember. Leptocerus sp., 5, June to November. Leptocerus dilutus, 7, May, August. Molanna uniophila, 5, July, August. Orthotrichia, 1, October. Platyphylax subfasciatus, 2, May, August. Neuroptera larve (1.7)— Sialis infumata (1.7), 75, all year. 312 INSECT LARVA}—Continued. Coleoptera larve— Carabus, 1, July. Parnid, 1, March. Pelocaris femoratus, 1, June. Hemiptera nymphs— Plea minutissima, 1, July. INSECT PUPA: (8), all year: Corethra punctipennis, 12, August, September. Chironomus sp. (3), 128, all year. C. decorus (3.5), 139, April to October C. digitatus, 2, July. C. fulviventris, 18, March to May. C. lobiferus, 26, March to November. C. viridis, 1, July. Palpomyia sp., 4, August. Probezzia glaber, 2, August, September. P. pallida, 1, August. Protenthes choreus, 22, April to June. Tanypus sp., 7, June. T. carneus, 4, May. Tanytarsus dives, 13, March, April. ADULT INSECTS (1.5), all year: Ammophila sp., 2, August, October. Aphodius inquinatus, 2, April, May. Brachonid sp., 1, July. Camponotus, 1, May, August. Carabid, 1, April. Chironomus sp., 13, April to November. C. plumosus, 2, April. Corethra punctipennis, 7, July, August. Collembolid, 1, July. Corixa, 40, February to November. Enchenopa binotata, 1, July. Heterocerus sp., 1, April. Lachnosterna, 1, May. Noctuid sp., 1, July. Platyphylas subfasciatus, 1, August. Sawfly, 1, April. Scarabezid, 1, April. ARaAcHNIDA, February to September: Arrhenurus, 7, May. Atax turgidus, 2, June. Limnesia, 15, May to July. Mites, unidentified, 17, February to Sep- tember. Spider, 1, June. AmpuiPopa (4.3), all year: Dikerogammarus fasciatus, 9, February to September. Gammarus limnzeus, 4, March. Hyalella azteca (4.1), 232, all year. BULLETIN OF THE BUREAU OF FISHERIES. Isopopa: Asellus communis, 4, February, March. Dercapopa, (1.2), May to August: Cambarus propinquus, 8, May to July. C. virilis, 1, July. Crayfish, unidentified, 16, May to August. ENTOMOSTRACA (21.4), all year: Cladocera (20)— Acroperus, 6, June, July. Bosmina, 7, July. Ceriodaphnia, 13, May to August. Chydorus sphericus, 29, February to October. Daphnia longispina hyalina (9.8), 215, all year. D. pulex (1.2), 42, June to December. D. retrocurva, 8, September. Diaphanosoma, 6, July, August. Eurycercus lamellatus (1.2), 95, all year. Leptodora, 135, all year. Pleuroxus procurvatus, 5, July, January. Copepoda (1.1), all year— Canthocamptus, 1, June. Cyclops albidus, 2, March. C. bicuspidatus (1), 86, all year. C. fuscus, 3, January. C. leuckarti, 2, August. C. viridis, 1, July. Diaptomus, 1, July. Nauplii, 2, June. Ostracoda, 52, all year. BRYOZOA: Statoblast, 1, September. Pectenella, 3, February, August. Mo..usca (2. 4), all year: Amnicola, 13, June to November. Campeloma, 10, May, June. Corneocyclas, 24, May to June. Limnza, 2, June, October. Physa heterostropha, 41, all year. Planorbis, 12, January to October. Snail eggs, 1, September. Sphezridz, 27, March to October. Spherium occidentale, 9, January to October. Valvata tricarinata, 6, May. OLIGOCHATA (1.4), February to September: Limnodrilus, 53, February to September. Tubifex, 2, June. Goropius, 23, December to July. HiruDInEA, January to March: Unidentified leech, 11, May, July. Glossiphonia stagnalis, 4, January to March. G. complanata, 2, June. Protozoa: Arcella, 1, April. HABITS OF YELLOW PERCH. 313 PLANTS (5.5), all year: PLANts—Continued. Unidentified remains (4), 225, all year. Alge, all year—Continued. Alga, all year— Spirogyra, 1, November. Unidentified, 7, April to September. Tabellaria, 1, November. Aphanothece, 19, April, September. Ceratophyllum, 1, January. Chara, 2, March, April. Elodea, 5, March, September. Closterium, 1, April. Lemna, 8, May to December. Desmids, 1, April. Plant leaves, 7, May to July. Diatoms, 20, October to March. Plant seeds, 12, May to March. Filamentous alge, 121, all year. Potamogeton, 3, May to October. Gelatinous alge, 4, July. Vallisneria, 1, October. Hydrodiction, 1, July. Wolffia, 2, June. Protococcus, 5, January, April. CaCO, CRYSTALS, 17, February to April.¢ Rivularia, 3, October. SILT AND DEBRIS (6.1), 276, all year. QUANTITY OF FOOD CONSUMED AND RATE OF DIGESTION. After the constituents of perch food had been ascertained and their percentages by volume determined, it became necessary, in order to gain some idea of a perch’s food requirements from day to day, to find out how much it could consume in a given time and how fast digestion progressed. With such purposes in mind, a medium-sized perch (weight, 48 g.; volume, 50 c. c.) was placed in a 5-gallon spherical glass aquarium and fed all it would eat from June 19 to July 20, 1916. The results of these experiments are shown in Table 13. Similar experiments were carried out later on smaller and larger fish and are in part summarized in Tables 11 to 15. The largest perch under observation were three individuals weighing about 247 g. and having volumes of about 250c.c. They were tested from December 18, 1916 to January 23, 1917, and ate only a few Dikerogammarus, although they were offered the same foods as smaller fishes tested at the same time (Tables 12 and 14). This agrees with the observations of Knauthe (1907), who stated that large carp usually would not eat when the temperature was below 6 to 8° C. As to the volume of the food in proportion to the bulk of the perch eating it, we have only a few observations. Table 13 shows that a perch displacing 50 c. c. of water ate the following percentages of its own volume per hour when given more than it consumed: Damselfly nymphs, 0.3 per cent; snails, 0; minnows, 0.46 per cent; earth- worms, 0.32 per cent. On January 12, 1917, a perch having a volume of 2.1 c. c. atea minnow (Pimephales notatus) which had a bulk of about 0.7 c. c. Reighard (1915, Pp. 237) gives instances where adult perch ate other individuals of their own species which were almost as large as themselves. The tables show that the same foods were digested more rapidly by small than by large perch and that, when fish of similar size ate at different temperatures, digestion was slower at lower temperatures. To take a concrete illustration: A perch about 62 mm. long ate seven chironomid larve (having a volume of 0.3 c. c.) at 2.5° C. and digested them in 43.7 hours; the same individual at 16° C. ate 0.84 c. c. of chironomid larve and digested them in 22 hours. A perch measuring 30 c. c. in volume ate 78 chironomid larve (having a volume of 4.2 c. c.), digesting them in 46.5 hours at 2.5° C. At 24° C. this perch ate 26 damselfly nymphs (no chironomid larve were available) having a volume aQn Nov. 23, 1917, calcium carbonate crystals were found in one of three perch caught at a depth of 2.5 m. in Lake Mendota. 314 BULLETIN OF THE BUREAU OF FISHERIES. of 2.5 c. c. and digested them in 23 hours. Taking asa basis for calculation Muttkowski’s (1918) estimate of the number of chironomid larve in shallow water and recent studies of deep-water fauna by Birge and Juday, there are about 474,750,000,000 in Lake Mendota. The number of perch they would support may be computed roughly: A perch of medium size if eating nothing but chironomid larve would average about 4.2 c. c., or 78 individuals of various sizes, per day, which would amount to about 1,533 c. ¢., or 28,470 individuals per year. On such a basis the chironomid larve could support 16,675,447 perch per year. Such methods of estimating are highly speculative at present but give some gross approximation as to the number of perch that might possibly live in Lake Mendota. After studies have been completed which are now being carried on by the Wisconsin Geological and Natural History Survey concerning the animal population of the various lake habitats and the chemical composition of animals which may serve as fish food, and after the writers have made more extensive experiments on the rate of consumption and digestion of different foods, it will be possible in a few years to speak with more authority ~ concerning the productive capacity of lakes. Piitter (1909) has made careful studies of the food requirements of the smelt and herring, which he expresses in terms of copepods. He states that the smelt needs the following numbers of copepods daily during its first season: May 6 to 29, 124 (1 mg.); May 29 to July 29, 248 (2 mg.); July 29 to September 25, 496 (4 mg.). The herring needs: July 13 to 30, 3,300 (26.6 mg.); July 30 to Sep- tember 20, 6,080 (49.9 mg.); September 20 to November 15, 4,470 (38.8 mg.). To keep in good condition the smelt would require 100 to 500 copepods daily during its growth period, and the herring 3,000 to 6,000. Little is known concerning details of digestive processes in fishes. Denis (1912) has measured the amount and composition of urine given off by sharks and goosefishes. Knauthe (1898) states that the amount of nitrogen given off increases as the temperature of a fish rises. Greene (1914) has made interesting studies of the utilization of fat by the salmon during its migration. Piitter (1909) found that a goldfish would change the contents of its intestine more often if peristalsis was artificially stimulated by suspending fine sand in the water. He also analyzed the substance in the carp and found it to contain: Water, 78.85 percent; dry substance, 21.16 per cent; albumen, 17.38 per cent; fat, 2.57 per cent; ash, 1.22 per cent; nitrogen, 2.91 per cent. Kmnauthe (1898) carried out extensive feeding experiments with carp. He states that, when no protein was fed, carp slowly became unable to digest pure starch, 10 times the usual amount of nitrogen being given off in the excreta. However, if protein was fed after a long period of exclu- sive carbohydrate diet, starch was again normally digested. Old fish did well when fed nothing but rice meal, and Knauthe believed this was because the proteins in the gonads were utilized. Fish died on an exclusive diet of lean meat. Ability to digest starch was lowered by deficiency of minerals in the diet. Ptitter (1909) states that it was impossible to rear smelt, herring, or carp when they were fed nothing but small crus- tacea. About the only generalization that can be made from the facts reviewed is that some variety is apparently necessary in fish diet. In the present work no essentially new contribution has been made to digestive processes in fishes, except for the points already reviewed relating to amounts consumed and the rate of digestion at different temperatures. These observations agree with those of Knauthe (1907) and Fibich (1905). One other interesting fact was noted. During HABITS OF YELLOW PERCH. 315 February, March, and April, 1916, the perch in Lake Mendota and the crappies in Lake Wingra had considerable amounts of beautifully regular calcium carbonate crystals (which were ‘usually embedded in a brownish, amorphous matrix) in their intestines. The food of 10 perch examined on each of the following dates contained the amount of crystals indicated: February 1, 0.5 per cent; March 1, 8.1 per cent; March 29, 12.3 per cent; April 14, 1.2 per cent; April 28, 2 per cent. At the same time the crappies (Pomoxis sparoides) in Lake Wingra showed the following amounts: March 11, 4 per cent; March 18, 0.2 per cent; April 22, 0.2 per cent; but the intestines of the perch in Lake Wingra contained none. ‘Two of the perch from Lake Mendota examined on March 29 contained 30 per cent of the calcium crystals. The remainder of the.food in one of these consisted of 45 per cent silt and bottom débris, 5 per cent chironomid larve, 1 per cent Corethra larve, 17 per cent plant remains, 2 per cent filamentous alge; in the other, of 35.9 per cent silt and débris, 30 per cent plant remains, 3 per cent Corethra larve, 1 per cent chironomid larve, o.1 per cent gelatinous alga. Birgeand Juday (1911, pp. 108, 171) analyzed the mud from the bottom and the crust from aquatic plants in Lake Mendota. The former contained, in percentages of dry weight, 33.21 per cent, and the latter, 47 per cent of calcium oxide. As has been previously stated, the perch feed largely inshore during February, March, and April, and the two individuals just cited, which showed a high percentage of crystals, also contained bottom mud and plant remains. ‘The crystals may be accounted for on the supposition that calcium carbonate taken in through the mouth is dissolved in the stomach and that crystals form in the intestine as water is withdrawn during absorption. VARIETY IN FOOD AND ADAPTABILITY IN FEEDING. Knauthe (1907) pointed out that certain fishes, such as the trout and the perch, changed readily from one type of available food to another; but others, like the pike, the smelt, and the lota, made such changes with difficulty and hence died more often during scarcity of certain foods. The senior writer (1918) developed this idea still further in his studies of the shore fishes of Wisconsin lakes and also demonstrated that different species of fishes manifest a rather marked degree of specificity in choosing food. Each species, even though it may be versatile, shows preferences for particular foods, and some kinds of fishes select from a very limited number of foods. Fishes certainly select specific foods from those available, and it is only by examining the contents of their alimentary canals that preferences can be determined. Abundant foods are often apparently avoided in one lake by a particular fish, but the same food is eagerly eaten by it in another. As Knauthe (1907) says: “The value of materials as food must be determined biologically even more than phenologically or chemically.” One example will illustrate the fact that different species show specific preferences. On April 22, 1916, six species of fish were caught at the same time and place in Lake Wingra, and an analysis of the food eaten gave the following results: Ten breams (Abramis chrysoleucas) had eaten of Chironomus decorus larve, 3.5 per cent; Chironomus sp.? larve, 2 per cent; Cricotopus trifasciatus larve, 2.6 per cent; mayfly nymphs, 2 per cent; chi- ronomid pup2, 1 per cent; Cricotopus trifasciatus pupe, 1.5 per cent; Hyalella, o.1 per cent; ostracods, 0.2 per cent; Canthocamptus, 1.5 per cent; Cyclops, 33.8 per cent; Daphnia pulex, 19.3 per cent; Chydorus sphericus, 5.1 per cent: Bosmina longirostris cornuta, 1.5 per cent; Physa, 2 per cent; Oscilla- toria, 4.7 per cent; alge, o.1 percent; flagellates, o.2 per cent; Volvox, 0.7 per cent; plant remains, 9 per cent; fine débris, 9.1 per cent. 110307°—21——21 316 BULLETIN OF THE BUREAU OF FISHERIES. Five crappies (Pomoxis sparoides) had consumed of Chironomus fulviventris larve, 24.2 per cent; mayfly nymphs, 13 per cent; Callibetis nymphs, 7.2 per cent; Enallagma hagenit nymphs, 6 per cent; chironomid pupe, 13 per cent; Corixa adults, 3.6 per cent; Hyalella, 3.2 per cent; ostracods, 1.3 per cent; Cyclops, 26 per cent; Daphnia, 2.6 per cent; Bosmina, 0.6 per cent; Eurycercus +; calcium carbonate crystals, o.2 per cent. Thirteen perch contained of fish eggs, 0.7 per cent; minnows, 8 per cent; fish remains, 7 per cent; insect larve, 1.5 per cent; Protenthes larve, 2 per cent; Chironomus decorus larve, 7.6 per cent; C. fulviventris larve, 38.2 per cent; Probezzia pallida larve, 0.5 per cent; caddisfly larve, 0.6 per cent; Callibetis nymphs, 7.8 per cent; Enallagma hageni nymph, 1.4 per cent; chironomid pupe, 6 per cent; Hyalella, 1.5 per cent; ostracods, +; Eurycercus, 0.1 per cent; Physa, 6.5 per cent; Planorbis, 0.2 per cent; Pleurococcus, +; Chara, 1.4 per cent; fine débris, 8 per cent. Though the perch is versatile, it selects preferred foods from the environment, and preferences apparently vary more or less at different ages, seasons, and localities. The staple articles of diet for adult perch throughout the year are chironomid larve and cladocerans, but with changing seasons there may be great variation in the pro- portions of either. Furthermore, perch of the same size caught at the same time and place have usually eaten the same kinds of food, but at times have not. As a rule the nature of the food indicates that it was secured on or near the bottom, but schools of perch are sometimes seen feeding at the surface of the lake. This is particularly true in the early morning. Judged by the success of line fishing, feeding is largely diurnal, for few perch can be caught at night. This indicates also that perch depend upon their sense of sight to a marked degree, and this view has been supported through the observation of individuals fed in glass aquaria. In order to discover whether adult perch ate different foods at various depths, comparisons have been made which include all catches in Lake Mendota. These are summarized in Tables 16 and 17. The first shows the percentage of food at different depths; the latter, the kinds of foods which exceeded all others in volume. Both tables show: (1) That food is more varied in shallow water and that it consists largely of chironomid larve, Corethra larve, Daphnia, Corneocyclas, and bottom mud in the deeper parts of the lake; (2) that the following foods decrease in amount eaten fn passing from shallow to deep water—small fishes, mites, adult insects, crayfishes, Hyalella, copepods, snails, leeches; (3) that the following foods increase in passing from shallow to deeper water—insect larve, insect pupze, bottom mud, Cladocera, small clams, oligochetes, plants; and (4) that in general the perch have eaten the foods which are most abundant at the depth where they are caught. This indicates that they do not change rapidly from one stratum to another; that is, there are usually no rapid vertical migrations. The fact that insect pupe, largely those of midges, are eaten mostly in deep water indicates that they are secured in the bottom mud before beginning their migration to the surface. Plants occur in greater amounts in perch from deeper water, and this is probably because remains of plants which have been washed loose and broken up by storms are so arranged as to be lying upon the bottom. Deposits of such plants are favorite resorts for insect larve. The tables just cited and evidence from many other sources show that the perch usually feed on or near the bottom. The contrast in the food of perch from different depths may, perhaps, best be indicated by the following specific instances where individuals were caught at the same time: HABITS OF YELLOW PERCH. 317 JULY 1, 1915. Depth, 18.3 m.; number examined, 9. Food—Chironomus decorus larve, 21 per cent; Protenthes choreus larve, 12.4 per cent; Corethra punctipennis larve, 15 per cent; Chironomus decorus pupe, 21.6 per cent; Corethra punctipennis pup, 0.5 per cent; mites, o.1 per cent; ostracods, o.r per cent; Daphnia hyalina, 13.3 per cent; Corneocyclas idahoensis, 8.3 per cent; Hyalella, 0.5 per cent; Gordius (from chironomid larvae), 6.4 per cent. Summary.—Midge larve, 48.4 per cent; midge pupz, 22.1 per cent; mites, 0.1 per cent; ostracods, 0.1 per cent; cladocerans, 13.3 per cent; clams, 8.3 per cent; amphipods, 0.5 per cent; Gordiacea, 6.4 per cent. Depth, 15 m.; number examined, 5. Food—Chironomus decorus larve, 14 per cent; Corethra puncti- pennis larve, 9.4 per cent; Protenthes choreus larve, 6 per cent; Chironomus decorus pupe, 33 pet cent; Protenthes choreus pupe, 7.6 per cent; Daphnia hyalina, to per cent; Sphaeriide, 14 per cent; Gordius, 4 per cent. Summary.—Insect larvee, 29.4 per cent; insect pup, 40.6 per cent; cladocerans, 10 per cent; clams, 14 per cent; Gordiacea, 4 per cent. 4 Depth, 4 m.; number examined, 5. Food—Minnows, 19.4 per cent; Chironomus unidentified larve, 5-6 per cent; Chironomus abbreviatus larve, 1 per cent; Corethra punctipennis larve, 2 per cent; Chironomus decorus pup, 1 per cent; Cambarus propinquus, 36 per cent; Hyalella, 9 per cent; Eurycercus lamellatus, t per cent; Physa heterostropha, 16 per cent; Planorbis, 0.4 per cent; leech, 9 per cent. Summary.—Fish, 19.4 per cent; midge larve, 8.6 per cent; midge pupz, 1 per cent; cray- fishes, 36 per cent; amphipods, 9 per cent; cladocerans, 1 per cent; snails, 16.4 per cent; leech, 9 per cent. MAY 12, 1916. Depth, o.5 m.; number examined, 5. Food—Sucker eggs, 21.8 per cent; Chironomus decorus larve, I per cent; C. fulviventris larve, 1.6 per cent; Enallagma antennatum nymphs, 3 per cent; Argia nymph, r per cent; Leptocerus ancylus larve, 8 per cent; Sialis infumata larve, 16.9 per cent; Hyalella azteca, 0.8 per cent; Physa heterostropha, 8 per cent; Nephelopsis obscura, 19 per cent; bud scale, 0.4 per cent; Lemna, 1.5 per cent; filamentous alge, 2 per cent; sand, 13 per cent; débris, 2 per cent. Summary.—Fish eggs, 21.8 per cent; insect larve, 31.5 per cent; amphipods, 0.8 per cent; snails, 8 per cent; leeches, 19 per cent; plants, 3.9 per cent; sand, 13 per cent; débris, 2 per cent. Depth, 4 m.; number examined, 2. Food—Perch eggs, 99.7 per cent; Chironomus decorus larve, 0.1 per cent; Sialis infumata larve, o.1 per cent. Depth, 7 m.; number examined, 3. Food—Chironomus decorus larve, 51.6 per cent; Protenthes choreus larve, 21.7 per cent; Corethra punctipennis larve, 4.7 per cent; Sialis infumata larve, 15 per cent; Ammnicola limosa, 1.3 per cent; leech, 5 per cent; fine mud, o.7 per cent. Summary.—Insect larve, 93 per cent; snails, 1.3 per cent; leeches, 5 per cent; mud, 0.7 per cent. Depth, 15 m.; number examined, 3. Food—Chironomus decorus larve, 8.3 per cent; Protenthes choreus larve, 8.3 per cent; Corethra punctipennis larve, 2 per cent; Leptocerus ancylus larve, 1.3 per cent; Sialis infumata larve, 5 per cent; oligochetes, 28.3 per cent; bottom mud, 46.7 per cent. Summary.—insect larve, 24.9 per cent; oligochetes, 28.3 per cent; bottom mud, 46.7 per cent. Depth, 17 m.; number examined, 3. Food—Chironomus decorus larve, 18.7 per cent; Protenthes choreus larve, 8.7 per cent; Corethra punctipennis larve, 1 per cent; oligochetes, 46.7 per cent; Corneocyclas idahoensis, 0.7 per cent; bottom mud, 24.3 per cent. Summary.—Insect larve, 28.4 per cent; oligochetes, 46.7 per cent; clams, o.7 per cent; mud, 24.3 per cent. Except for the food of a few young perch and for the comparison (p. 319) between the examinations in Lakes Wingra and Mendota, no particular studies have been made of the perch from different lakes. It is probable that the feeding habits are rather uniform, but the food varies according to conditions in different localities. 318 BULLETIN OF THE BUREAU OF FISHERIES. VARIATION IN FOOD AT DIFFERENT AGES. In order to determine what foods were eaten by young perch during the first summer after they hatched, collections were made with a minnow seine at intervals during 1916 in shallow water east of the base of Picnic Point in Lake Mendota. The results of the food examinations are summarized in Table 18. The perch were very uniform as to size, and it will be noted that the average length showed a regular increase as the season advanced. The table shows that Cyclops, other small crustaceans, and minute insect larve are replaced to a large extent by Hyalella and good-sized insect larve as the perch increase in size. To compare the food of the perch summarized in Table 18 with the food of those from another place in Lake Mendota on a date close to one of those utilized in the usual locality, 10 small perch were collected from the mouth of Six Mile Creek on August 8, 1916. They had eaten of Tanypus monilis larve, 2.5 per cent; Chironomous lobiferus larve, 7.8 per cent; mayfly nymphs, 18.5 per cent; Betis nymphs, 8 per cent; Cenis diminuta nymphs, 23.4 per cent; Corixa nymphs, 16.3 per cent; Chironomus lobiferus pup, 13.1 per cent; Hyalella azteca, 10.3 per cent; ostracods, +. If these results be compared with those for August 7, in Table 18, it is apparent that only three of the same items have been eaten in the two localities, yet there is general similarity. About the same types of foods are eaten in about the same proportions. Through the kindness of A. R. Cahn we were able to examine small perch from Oconomowoc Lake. Though the individuals were more variable in size than those examined in Lake Mendota, the same food changes are evident (Table 19). Small insect larvee and entomostracans are succeeded by larger larvee and Hyalella. If these two tables showing the food of small perch are compared with Table 9, which gives the results for adults, it is evident that at the close of the first summer the food of the young has become like that of adults. RATE OF GROWTH ON DIFFERENT FOODS. To be of most significance, the determinations of the rate of growth should be made on perch of various ages at different temperatures. Kmnauthe (1898) performed experiments which indicated that metabolism is more rapid in young fish than in old and that more protein food is necessary during youth. Older fish need apparently more mineral than young. Piitter (1909) says that the smelt and herring require nearly twice as much food after growing for a month. He found that a carp after two summers weighed about 500 g. and that it would increase to 1,250 g. by the middle of the next August. In this paper it has already been shown that digestion is more rapid in perch at higher temperatures (p. 313). We have been able to test the rate of growth in perch of one size only and at one temperature. From August 19 to September 18, 1916, when the temperature of the water varied from 20 to 16.8° C., 26 small perch were placed in separate glass jars having a capacity of 4 liters each, and in lots of 3 were fed, as follows: : 1. Fish liver and flour mashed and mixed together. 2. Hyalella azteca alive. F 3- Plankton fresh from Lake Mendota. It consisted of Daphnia, 95 per cent; Leptodora, 4.5 per cent; alge, chiefly Lyngbya, 0.5 per cent. 4. Earthworms alive. HABITS OF YELLOW PERCH. 319 . Insects—Corixa, Plea nymphs, Notonecta nymphs, damselfly adults, crickets, midges. . Chironomus decorus larve alive. . Fish cut into small pieces. . “Normal’’ diet, consisting of all the kinds of food fed under 1 to 6, but no single one in large enough quantity to give complete satisfaction by itself. The two jars in which these fish were kept contained also Elodea. g. Starved. ona an The fish ate all classes of food readily except the insects. Although Corixa and other varieties which occurred in perch of similar size in nature were offered, they were never taken in any quantity and were often refused altogether. The practice with all the foods was to change the water each morning and in midafternoon to add a fresh supply of food, which exceeded what might be eaten before the next day. The results of these experiments are given in Table 20. The foods would come in the following order, as judged by the rate of gain in weight and volume: Earthworms, Entomostraca, chironomid larve, amphipods, fish, ‘‘normal,” liver, and flour. The three perch fed adult insects lost almost as much as those which had nothing. It will be noted that there is no correlation between the gains in weight and volume. It is difficult to understand why the three “‘normal” individuals which were fed a variety did not gain as much as others which received only one kind of food during the entire month. Perhaps the extra energy required to digest a variety more than compensated for the diversity of chemical substances obtained. COMPARISON OF FOOD OF PERCH IN LAKE MENDOTA AND IN LAKE WINGRA. The fact that perch are individually smaller in Lake Wingra than in Lake Mendota is probably due to a number of causes, but one would naturally turn first to differences in food for an explanation of such variance. In Table 21 the various foods eaten by the perch in each lake is given by months. The averages show that fish, insect larve, insect pupz, adult insects, isopods, and copepods are eaten in greater amounts in Lake Wingra than in Lake Mendota; the opposite is true of mites, crayfishes, amphipods, ostracods, cladocerans, snails, clams, leeches, oligochztes, plants, silt and débris, and CaCo, crystals. In all but two months in Lake Wingra, insects as larve, pupe, or adults form half, or more than half, of the food. In Lake Mendota the months are equally divided, as regards the particular foods eaten in maximum amounts, between Cladocera and insect larve. An examination of figures 6 to 29 will show many other minor differences in details between the two lakes. Among the insects the chironomids do not differ much, but Wingra excels in chironomid pupe and in odonate and mayfly nymphs. Among the cladocerans the amount.of Leptodora was about the same in the two lakes (4.6 to 4.7 per cent); Daphnia was in excess in Mendota (20.9 to 4.8 per cent), and Eurycercus in Lake Wingra (0.1 to 30.3 per cent). Another difference between the perch in the lakes under consideration is shown in Table 22, which demonstrates that there are two seasons in Wingra when many of the perch have little or no food in them and only one in Mendota. The empty perch in April are doubtless due to the neglect of feeding on account of breeding. The fasting period in Wingra during August and September has no counterpart in Mendota and is equally characteristic of both sexes. It is probably due to the continued high tempera- ture, from which there is no escape, as there is in Mendota, and to the extreme turbidity 320 BULLETIN OF THE BUREAU OF FISHERIES. of the water. In late summer the water in Lake Wingra is murky with myriads of alge. ‘The perch are pale in color and apparently in poor condition. The chief respect in which the perch of Wingra differ from those in Mendota, in regard to food, is (1) that they eat more insect larve and less of entomostracans, (2) that through most of the year they apparently feed more among water plants, and (3) that they have longer periods when little or no food is eaten. COMPARISON OF FOOD OF PERCH AND CRAPPIE IN LAKE WINGRA. From February to November, 1916, the food of crappies (Pomoxis sparoides) from Lake Wingra was studied, and the results for the nine months may be compared with those for perch. The total percentages of foods eaten by both was as follows, the perch being placed first in each case: Fish, 12.7 to 8.8; insect larve, 52.8 to 25.5; insect pupz, 11.4 to 7.9; adult insects, 1.3 to 4.8; mites, 2 to +; amphipods, 0.3 to +; clams, 0.05 to 0; leeches, 0.2 to +; oligochates, 0.1 to 0; plants, 3 to 0.4; débris, 1.2 tox. In other words, the perch eats more of fish, insect larve, insect pupz, mites, amphipods, snails, clams, leeches, oligochztes, plants and débris; the crappie more of adult insects, ostracods, copepods, cladocerans. These proportions clearly indicate that perch feed largely on or near the bottom, while crappies hunt more toward the surface and among water plants. The structure of the crappie is more specialized than that of the perch and would indicate greater adaptation to particular conditions. Its mouth is more upturned, suggesting feeding toward the surface of the water, and the body is more compressed, indicating a habitat among aquatic vegetation. A full account of the observations on the crappies of Lake Wingra has already been published (Pearse, 1919). RESPIRATION. A perch must obtain the oxygen necessary for its metabolic activities from the water in which it lives. Water free to absorb gases from the atmosphere will contain about 35 per cent oxygen, 65 per cent nitrogen, and a trace of carbon dioxide. The total amount of gas which may be absorbed varies with the temperature of the water. At 0° C. aliter of water, when the pressure is 760 mm., can absorb 41.14 c. c. of oxygen, 1,796.7 ¢. c. of carbon dioxide, and 20.35 c. c. of nitrogen. At 20° C. the amounts will be: Oxygen, 28.38 c. ¢.; carbon dioxide, 901.4 c. c.; and nitrogen, 14.03 c. c.; at the boiling point of water none of the gases will be absorbed. When there are many living plants present, the amount of oxygen may rise above the saturation point; when oxygen is used up (decomposition, respiration, etc.), it may fall much below saturation or even be absent altogether. A perch, then, normally lives in water which may vary greatly in its gaseous content at different seasons. Supersaturation with oxygen appears to offer no particular difficulties for fish, but when this gas is scanty there may be trouble in obtaining a sufficient supply for respiratory needs; yet some species are able to live in water containing a very small amount. Winterstein (1908) states that 0.7 c. c. of oxygen per liter was enough to sustain life in Leuciscus erythropthalmus, but when the amount was decreased as low as 0.4 to 0.5 c. c., death ensued. Most fishes show signs of distress when the oxygen is 1 to 4c. c. per liter. In natural bodies of water the carbon dioxide usually increases as the oxygen decreases, but in amounts such as occur in nature its presence does not HABITS OF YELLOW PERCH. 321 appear to be particularly detrimental, provided enough oxygen is also present. Winter- stein (1908), however, thinks that fishes are affected by the presence of carbon dioxide, becatise some species succumb when its tension is only 8 to 12 per cent of the total pressure of gases; but he also found that in two instances as much as 144.7 and 204.6 c. c. per liter were required to overcome Leuciscus. Shelford and Allee (1913) conclude that the narcotic effect of carbon dioxide is more important for fishes than its action as an acid. Various species tested by them were affected injuriously when the amount present was from 5 to 37.5 c. c. per liter. As a rule, four factors are of chief importance for the normal respiration of fishes: (1) Sufficient oxygen for metabolism, (2) lack of enough carbon dioxide to be injurious, (3) favorable temperature, and (4) proper reaction (salinity or acidity) of water. Though oxygen and carbon dioxide are the only gases which usually affect the respira- tory activities of fishes, others may be of some importance at times. Methane and ammonia sometimes occur in certain restricted localities, and are injurious; nitrogen may, if present in unusual amount, give rise to the gas disease (Marsh and Gorham, 1905). But troubles from such gases are of rare occurrence. Gardner and Leetham (1914) have shown that a trout uses twice as much oxygen for respiration if the tem- perature of the water about it is raised from ro to 20° C. Wells (1913) found that fish died more quickly in alkaline than in acid water when gas conditions were poor. Marsh (1910) asserts that fish will not live in well-aerated distilled water and that they are very susceptible to dilute solutions of mineral acids. The respiration of fishes, then, requires reasonably pure water of proper chemical reaction and with a sufficient supply of oxygen. Tlie experiments of Shelford and Allee (1913, 1913@) have demonstrated that fishes respond to the conditions in their environ- ment in such a way as to spend most of their time in the optimum. Fishes are able to discriminate variations in the gas content of the water, and when placed in a graded series usually spend the most time where conditions are best. They are apparently more stimulated to turn away from unfavorable regions by the presence or carbon dioxide than by deficiency in oxygen, some turning back upon encountering 1.5 c. c. of carbon dioxide per liter. “‘We have in the experiments good evidence that fishes turn back from waters high in carbon dioxide and low in oxygen with precision and vigor. Also that if they enter such localities, they can not behave normally and may soon die.’ Wells (1913) also has shown that fishes are most active when in water containing a scanty supply of oxygen and has demonstrated (1915) that a number of fresh-water species select slightly acid water in preference to that which is alkaline. He also asserts (1913) that the future will show that the reactions of fishes are of more importance than their resistance to unfavorable conditions. Death after reaching the vital limit is unusual, but the avoidance of conditions which may mean death is of frequent occurrence. The behavior of fishes is such that it would usually keep them in optimum conditions; yet Juday and Wagner (1908) found that lake trout commonly entered deep waters which contained so little oxygen that they could not live in them for any length of time. Paton (1902) also observed that brook trout which were kept in water containing very little oxygen were able to survive for some time by remaining inactive on the bottom and thus reducing their metabolism to a minimum. ‘This brings us to the resistance of fishes to a marked deficiency in oxygen or to an unusually large amount of carbon dioxide. a Shelford and Allee, 1913, p. 251. 322 BULLETIN OF THE BUREAU OF FISHERIES. Packard (1905, 1907, 1908) kept top minnows, Fundulus heteroclitus, in oxygen-free water and found that they were able to live about three hours. He believes that these fishes must get oxygen from other sources than the atmosphere. His experiments support Mathews’ (1905) theory of respiration, which supposes that the oxygen of the atmosphere acts as a depolarizer and combines with nascent hydrogen produced during metabolism. If oxygen can be replaced by some other substance which will neutralize hydrogen, its presence is not necessary. Packard found that he could prolong the life of Fundulus in water which contained no oxygen by injecting carbohydrates into the body cavity. Wells (1913) determined that abundant oxygen and carbon dioxide was less injurious to fishes than a very small amount of both gases or than much carbon dioxide and little oxygen. He says (p. 345): “Oxygen in large amounts (10 ¢. c. per liter) antagonizes the detrimental effects of high carbon dioxide (50 c. c. per liter).” He also found that the most active fishes succumbed to a large amount of carbon dioxide before more slug- gish individuals, and that oxygen deficiency was more quickly fatal when the water was alkaline than when it had an acid reaction. The observations just reviewed, then, indicate (1) that fish are able to live for some time in water without oxygen; (2) that lack of oxygen is generally more injurious than excess of carbon dioxide; and (3) that gas conditions unfavorable for respiration are more quickly fatal in water with an alkaline reaction. The general resistance which fishes show to suffocation is in many species assisted materially by the use of the swim bladder as a storage reservoir for oxygen. ‘This organ apparently serves various functions and in different fishes may be used as a lung, an organ for making sounds, a hydrostatic organ, and as a respiratory reservoir. Woodland (r911) proved the hydrostatic function by weighing fish after subjecting them to differ- ent pressures. The use of the bladder as a storage reservoir for oxygen has been the subject of a number of investigations. Tower (1902), for example, studied a number of miatine fishes and found there were three gases present—oxygen, carbon dioxide, and nitrogen; that the amount of carbon dioxide might increase a little (0.25 per cent) during suffocation, but that it was usually 0.06 to 0.1 per cent of the total gases; and that the deeper the water from which fishes were taken, the higher the proportion of oxygen (in some fishes captured at considerable depths the gas in the bladder was practically all oxygen). Bridge (1891) showed that the secretion of gases into the swim bladder was under the control of the nervous system; and he found that there was usually an increase in the amount of carbon dioxide when the fish died of suffocation. The Cambridge Natural History states that in general the amount of oxygen in the bladder is less in fresh-water fishes than in those from the ocean. In fishes like the perch, in which the swim bladder has no duct connecting it with the outside, its functions are confined to storing reserve oxygen and regulating the specific gravity of the body (hydrostatic function). Finally, the density of the water in which fishes live affects respiratory activities. As pressure grows greater on account of increase in depth, the ability of the water to absorb gases is also increased. Furthermore, the comparative “hardness” or ‘‘softness’’ of water not only affects the density, but has a marked influence on the pressure of the gases present in solution. Sumner (1906) asserts that the membranes of fresh-water fishes are highly adapted to resist changes in density. There is an “irreducible minimum”’ of salts in the blood which is not released even when HABITS OF YELLOW PERCH. 323 the surrounding water has a very low salt content. Garrey (1905) found the osmotic pressure of blood of marine fishes was about half that of sea water. With this brief review of facts gleaned from literature on the respiration of fishes, the discussion will now be turned to the discoveries made during the present investiga- tions in regard to the respiratory activities of the perch in Lake Mendota. During the summer of 1915 it was noticed that many fish caught in gill nets in Lake Mendota were dead, and that there was a greater mortality in the region of the thermocline and below it than above (fig. 31; Table 2). Of 2,194 perch caught, 343 were dead and 1,154 alive above the thermocline; 610 dead and 87 alive below. This ‘101724 317 '14'21'28 4 ‘NN I8'25° 2° 4° JULY AUG. SEPT. OCT. Alive Dead 100. °° Fic. 31.—Percentage of perch caught alive or dead in gill nets above and below the thermocline, Lake Mendota, sors. , above thermocline; ------ , below thermocline. indicated that perch commonly entered water which contained too little oxygen for respiration. During the summer of 1916 careful observations were again made, and the same results were obtained. Though perch were usually most abundant imme- diately above the thermocline, large catches often occurred just below it, where there was no oxygen. The next problem was to discover how long perch could live in the oxygen-free water below the thermocline. Accordingly, from August 30 to September 4, 1916, when the amount of oxygen at a depth of 13.5 m. was 0.05 c. c. per liter and the carbon dioxide was 5 c. c. per liter, perch were let down into the stagnant region and left for various periods of time. The results of the experiments are shown as follows: 324 BULLETIN OF THE BUREAU OF FISHERIES. Perch Perch Hours submerged. Perch. | survive Hours submerged. Perch surviv- used. 5 used. Hi ing. ing. ao 8h ogiicosabvecdbecconenta ses snsesio 6 Oy] | a ead ea aretatere visite ict teteretste cies heals ieia ei ciotatara’s 6 2 DLO TEM ec cdisicsascciet atatinlewe tine tobias tere 2 Bull) Bixee hietaiatelelejalb vleietele a G Rib i Ge mites Bri athinld ote /diaietes fe 6 ° FS Sine wai = apclos oo carotene Bonn Aci) 6 | be enh Stes fone SAEAG. 1ACOSDeC ACAD HG 6 ° It will be observed that some perch lived for over two hours, but that none sur- vived for three. This suggested that perch might be able to enter the lower waters of the lake with impunity to take advantage of the abundant food supply there, coming up above the thermocline at intervals to breathe. This possibility made it necesary to observe the behavior of perch in water from below the thermocline. On September 4 water was pumped up from 13.5 m. (oxygen, 0.06 c. c. per liter; carbon dioxide, 5 c. ¢c.) into a large aquarium on the deck of a boat. Samples taken from the water in the aquarium showed that it contained about 0.3 c. c. of oxygen and 5 c. c. of carbon dioxide per liter. Nine fish, caught in a gill net half an hour before at a depth of 11.6 m., were placed in the aquarium and their behavior observed for two hours. They were compared at intervals with perch caught at the same time and place, which were kept in a large cage at the surface of the lake. One of the fish in the stagnant water turned on its side and became immobile (except for respiratory movements) after four minutes. After half an hour several were lying on their sides, but after an hour and forty minutes one individual was still right side up and, though inactive, appeared to be in normal condition. On the following day two perch, caught an hour before at 11.7 m., were again placed in water pumped from 13.5 m. into a glass aquarium, and their behavior was observed for an hour. Both of these individuals turned belly up within a few seconds; one floated at the top of the cage, but the other at times was at the bottom and at times at the top. At intervals both righted themselves, moved the fins, and wiggled about actively. After being in the stagnant water for an hour both were taken out and placed in the lake just below the surface. Half an hour later both had recovered, were right side up, and apparently in good condition. Two hours later both were released and swam away. During the time the perch were in the stagnant water the rapidity of their respiratory movements was observed, and the results, showing the number of respiratory movements per minute, are given as follows: Time. No. 1. | No. 2. Time. No. 1. | No. 2. 32 28 25 28 21 33 20 3r One individual stopped making respiratory movements for over two minutes in the midst of the experiment. In general, the rate of respiration decreased, but when the fish were placed in the lake again the rate rapidly increased. The rate of those HABITS OF YELLOW PERCH. 325 in the stagnant water was considerably less than that of those observed simultaneously in a cage in the open lake. But the water from 13.5 m. was cool (15° C.), and it would therefore be expected that the perch would respire at a slower rate than when in surface water (21.6° C.). In order to determine the normal rate of respiratory movements at different temperatures two healthy perch were observed in the laboratory on November 29, and the results are summarized as follows: Menipetatire (degrees/cetitigrade):. [050.100 /SCa 2 Ck AREA Aas ee 8 16 PA GIMET MADER ICE Lem cee ace Matai. sie ros oan wae hin smth ene labiensone spear nena shaS 24.1 38. 5 20 45-6 25 59°3 The data indicate that the rapidity of the respiratory movements of the fish placed in stagnant water was not materially increased or decreased by such treatment. Fro. 32.—Collecting tube and method of collecting gas from perch swim bladders. j, jar full of stagnant water continually pumped from depths of lake; m, small vessel for collecting mercury; ?, collecting tube filled with mercury. ‘The experiments described show that perch in Lake Mendota commonly enter the stagnant water below the thermocline and that they may remain there for an hour or two without suffocating. These facts suggested that perch might make use of the oxygen in the swim bladder while in the stagnant areas, and experiments were performed which showed the supposition to be correct. From August 23 to September 24, 1916, fishes were lowered in wire cages to depths varying from 12 to 13.5 m. The amount of 326 BULLETIN OF THE BUREAU OF FISHERIES. oxygen at such depths was about 0.05 c. c. per liter and the carbon dioxide, 5 c. c. per liter. The perch were allowed to remain for varying lengths of time in the stagnant water and then were pulled quickly (10 seconds) to the surface and placed in a large jar of water pumped from the same depth at which they had been submerged. They were opened as soon as possible under water and the gas in the bladder siphoned out into a mercury-filled collecting tube (fig. 32) of 5 c. c. capacity. The samples were carried into the laboratory in the collecting tubes and analyzed with a Haldane apparatus, the oxygen being absorbed with 10 per cent alkaline pyrogallol and the carbon dioxide with Io per cent potassium hydroxide. Every time samples were taken from perch which had been submerged in deep water, two or three control individuals which had been in a fish car at the surface were also tested. The details of the results of the analyses are shown in Table 23, and a summary is given in Table 24. Though considerable variation is shown, the latter table indicates that the oxygen in the bladder was used up while the perch were in the stagnant water, but the carbon dioxide did not increase. From the studies in Lake Mendota the following facts have been ascertained: (1) Perch commonly go into the stagnant water below the thermocline, where there is only a fraction of 1 per cent of oxygen per liter; (2) they may remain there for two hours or more without suffocating, but it is doubtful if they would feed for more than a few minutes; (3) when perch invade water which does not contain sufficient oxygen for respiration they apparently draw to some extent on the reserve in the swim bladder; (4) if through the action of the bladder as a hydrostatic organ a perch is adjusted to pressure conditions above the thermocline, it will, if it invades the lower regions where it is overcome by lack of oxygen and excess of carbon dioxide, tend to float up into levels where gas conditions are more favorable. REPRODUCTION. It has already been pointed out that the perch in Lake Wingra are generally of smaller size than those in Lake Mendota. This backwardness in growth, however, does not appear to retard the attainment of sexual maturity (Table 25). Judging by the measurements made on individuals from a school of young perch which remained near the base of Picnic Point (fig. 2) during the summer of 1916 (Table 18), and by obser- vations on the gonads of half-grown perch at various seasons, the authors believe that perch may become sexually mature in Lake Mendota at the end of two years of growth. Meek (1916), speaking of conditions in Europe, says: ‘“‘The perch appears to become mature when it is three years old.”’ After a perch attains sexual maturity the gonads in both sexes pass through a regular cycle of seasonal changes. After spawning is completed, the gonads remain small until late summer and then increase very rapidly in size for a month or more... By September they are almost as large asin thespring. The growth of the gonads, then, takes place for the most part in the summer, when food is most abundant, and there is little change in size during the winter months. By November, perch caught in deep water (18 to 20 m.) will often shed eggs when brought to the surface. Such individuals are, of course, not completely “ripe” but emit eggs on account of the decrease in pressure. Prof. C. L. Turner (1919) has made a careful study of the volumetric and cytological changes in perch gonads at Milwaukee, and his paper gives detailed information concerning the annual reproductive cycle. HABITS OF YELLOW PERCH. 327 The spawning season in Lake Wingra comes earlier than in Lake Mendota. This is due largely to the fact that the ice goes out sooner and the smaller volume of water warms up more rapidly. Figures 3 to 6 show that the period of activity associated with the migration of perch into shallow water for spawning came nearly a month earlier in Lake Wingra. In the spring of 1916 the ice left Lake Wingra March 20 and Lake Mendota April 8. The temperature just below the surface on April 18 was 10.6° C. in Lake Wingra and 4.6° C. in Lake Mendota. Our observations agree with those of Forbes and Richardson (1908), who state that the spawning of perch takes place in April and May, when the temperature of the water is 7 to 10° C.% Compared with other species of fishes and amphibians which lay eggs in the spring, the perch spawn rather early. During the spring of 1916 the following sequence was observed in Lake Mendota: April 12, the swamp-tree frog, Corophilus nigritus, was singing in the swamps along shore; April 20 to May 7, perch were spawning; May 2, the larve of the orl fly, Szalis inju- mata, were migrating on shore; May 12, suckers, Catostomus commersonit, were spawning ; May 30, crappies and dogfishes were frequenting bare spots alongshore; and June 19, crappiesin Lake Wingra werespawning. When most of the perch in Lake Mendota were spawning, the majority of those in Lake Wingra were already spent. In the autumn also the gonads of the perch in the smaller lake were noticeably earlier in reaching the large size characteristic of the cooler months, and this is again correlated with the earlier cooling of the lake. Perch come into shallow water alongshore to breed. The males precede the females and remain longer on the spawning grounds. This means that there are many more males than females in shallow water from the middle of April until the early part of May.?. For example, on April 28, 1916, a 1-inch mesh gill net, pulled from a depth of less than 3 m.,contained 380 perch, and all but four wereripe males (Table 26). Therewere three ripe females and one immature male. On May 2 and 12 there was still a great preponderance of males in the nets set in shallow water, but on later dates the sexes became more or less similarly distributed at all depths. The males evidently came inshore and remained during the entire spawning season; the females left deep water for only a short time to lay their eggs. Meek (1916, p. 281) records similar behavior for the plaice: ‘‘Results appear to show that the males appear first at the spawning ground and remain during the season, whereas the females depart shortly after the ova are shed.”” Abbott (1878) states that perch go in pairs to the spawning beds. In our gill-net catches a ripe female was often surrounded by several males. This indicates that a female may be attended by more than one male. Breeding instincts appear to dominate feeding instincts at the time of spawning. Table 22 shows that about 6 per cent of the individuals captured during the breeding season contained no food, and it was mostly the males that were empty. Another difference in feeding activities was noted between the sexes. The fishermen on Lake Mendota have stated on various occasions that they always caught more females than males when fishing in deep water through the ice with hook and line. The following observations support this view: December 29, 1916, 13 m., 40 females, o males; Decem- @ On May 7, 1920, at 4.25 p. m., the writer set two 4 by 7s gill nets, tied end to end, at a depth of 2.7 m. on the south shure of Lake Mendota. At 10.45 8. m. on May 8 the 1!4 inch mesh net contained a rock bass anda pickerel. The r-inch mesh net at 8.15 a. m. on May 8 had caught 921 ripe male perch, ro ripe female perch, 9 spent female perch, 4 female perch which had whitish eggs in their ovaries, and 1 mud puppy. ‘The food of the last consisted of crayfishes, 92; Physa heterostropha, 4; plant remains, 2; Leptocella larva and case, 1; perch eggs, 1. The water temperature (first figure in each set indicating depth in meters; second, degrees centigrade) was as follows: 0, 11; 1, 10.8; 3, 10.4; 4, 10; 5, 9-7; 6, 8.1; 7, 7-2; 8, 7.1; 10, 6.9; 13, 6.5; 15, 6.4; 18, 6.1; 20, 5.8; 23.5, 5.6. 328 BULLETIN OF THE BUREAU OF FISHERIES. ber 30, same place, 38 females, 2 males; January 4, 1917, 15 m., 5 females, 3 males; January 6, same place, 9 females, 7 males; January 8, same place, 34 females, 7 males; January 25, 18 m., 28 females, 10 males. These facts indicate that the females feed more actively during the winter or that they exceed the males in numbers in the deeper parts of the lake. The egg string deposited by a perch which had been kept for several months in a running-water aquarium is shown in Plate LXXXIII, figure 2. This contained 2,650 eggs and was deposited on May 1, 1916, when the water temperature was 12° C. A string was also laid by another individual in the same aquarium on April 18. Forbes and Richardson (1908) mention a string recorded in one of the laboratories of this Bureau which measured 88 inches in length and weighed 41 ounces after the water had been drained from it. The strings swell very rapidly and harden somewhat after leaving the body of the female. They are often thrown over stones, plants, or other objects in the water. Gorham (1912) states that they may be attached to willow roots. The same authority says that eggs hatch in 8 to 10 days and that the small fry hide in nooks alongshore until they appear in schools as fingerlings. Hankinson (1908) and Reighard (1915) mention seeing schools of small perch in shallow water, and the latter notes that there may be small fishes of other species with them. On August 23, 1916, a school of about a thousand young perch was observed near our dock just north of the University of Wisconsin, and it remained in that locality for over a week. Meek (1916, p. 290) says of the European perch: ‘‘The larva measures about 5 mm. when hatched, and in the course of a year the young attains a length of 6 cm., and in two years about 13 cm.’’ We have already noted that perch hatched in the spring of 1916 in Lake Mendota had attained by August 30 a length of 68 mm., without the tail fin, and that perch less than 130 mm. long were sexually mature (Table 25). MIGRATIONS. The perch in lakes frequent various localities at different times. In general, migrations are correlated with the changes accompanying the rhythmical sequence of day and night or with those associated with seasonal succession. In order to secure data on the numbers of perch at different depths, fishing was carried on simultaneously at various levels. The catch per hour in a gill net gives a fair idea of the number of perch present, but rather wide seasonal variations are to be expected. Fishes will not be captured unless they are moving, and the lesser activity accompanying lower temperatures will cause smaller catches. Another means the authors have used for judging the number of fishes present in any locality is by the catch per hour with hook and line. During stormy weather the number of perch secured in Lake Wingra from a drifting rowboat with hook and line often exceeded that taken in gill nets, probably because the shallowness of the lake made it inexpedient for the fishes to move about much in windy weather. Neither gill nets nor hooks give accurate data as to the actual numbers present. However, they do give information which is of value for judging comparative numbers when they are used simultaneously at various depths. An examination of Tables 2 to 5 and 27 shows several points of interest in regard to the abundance of fishes at various seasons, and it is possible to make a number of generalizations from the data presented. In Lake Mendota the course of the annual migration is pretty definite. In the winter most of the perch are in deep water. As soon as the lake is free from ice there is a migration inshore for spawning, but the perch HABITS OF YELLOW PERCH. 329 soon return to deep water and remain there until the lack of oxygen drives them into shallower regions. As soon as the autumnal overturn renews the oxygen the perch return for the most part to the depths of the lake. Less marked migrations of the same general type also take place in Lake Wingra, but there is no stagnation period in the summer. Figures 33 and 34 bring out a point which has already been discussed to some extent under food and respiration; that is, though the perch are obliged to live above the thermocline from August to October, they descend at intervals into the cool, stagnant region below, probably to take advantage of the abundant food there. Wells (1915) has pointed out that fishes generally prefer water which has a slightly acid reaction to : @ a ae a WT eel) ae ENE SLR CO 123 a A fa 9 ED 19 AP 2 feogod | fe | | | BRSERHT ith nae hie a te EEE EEE EEE EEE EEE Fic. 33.—Perch caught in gill nets set at various depths, Lake Mendota, 1915. The curve indicates the thermocline. All nets used were r-inch bar mesh. * indicates a net 4 by so feet which caught only one-fifth as many perch as the other nets used, which measured 3 by 75 feet. Nets were left in the water about 24 hours. ‘The ice left the lake April 9 to rz; the fall over- turn took place October 9 and ro. that which is neutral or alkaline. Of course, such behavior would tend to keep perch in deep water or near the bottom vegetation. Gurley (1902) is an ardent advocate of temperature as the controlling factor in the seasonal migrations of fishes, but in Lake Mendota it can have but slight influence. The perch come into shallow water in spring, when the temperature is low, uniform at all depths, and the same as that which has prevailed for several months; in autumn they descend into deep water when the tempera- tures are again uniform throughout the lake. The food and the net and line catches both indicate that the perch remain on or near the bottom and in as deep water as possible throughout the year. The migrations into shallow water are to spawn and to escape stagnant conditions during the summer. 330 BULLETIN OF THE BUREAU OF FISHERIES. One other possibility remained to be tested, however. During the period of stag- nation in the lower water the perch might remain on the bottom in the region of the thermocline or spread out over the whole lake to feed on the plankton organisms in the water containing oxygen. ‘The latter alternative seemed improbable from the fact that it is easiest to catch perch near the bottom at any season, but it was decided to perform an experiment to find out. Accordingly, on August 10, 1916, four 1-inch mesh, 3 by 60 feet, gill nets were set north of the University of Wisconsin in Lake Mendota. At this time the thermocline was well established at a depth of 9 meters, and the gaseous content of the water (according to titrations by the Winkler method; Birge and Juday, 1911) at certain depths was as follows: At 18 m.—Oxygen, o.o1 c. c. per liter; carbon dioxide, fk Ecco En RS — rT |_|s3 le | ae | |?s| Bim Clim i] | fey fT | a or | oe ae HESS SSa- BL) NDEBELE As Re RgSS2 =a DSO S sO eee eee eee eee ee Fic. 34.—Perch caught in 3 by 60 feet, 1-inch bar mesh gill nets, Lake Mendota, 1916. Nets were left in the water for various periods of time, but those set on any particular day were left for the same length of time, and the catches for that day at different depths are therefore comparable. The curve represents the thermocline. ‘The ice left the lake on April 8; the fall overturn occurred October 5 to 10, f indicates 10 perch caught at 16 m. in about 2 minutes while washing net. 10.31 c. c. At 14.5 m.—Oxygen, o.o1 c. c.; carbon dioxide, 4.1 ec. c. At 13 m.— Oxygen, 0.02 c. c.; carbon dioxide, 4.17 c.c. At 6.6 m—Oxygen, 4.49 c. c.; carbon dioxide, oc. c. One net was set at 19.2 m. on the bottom; another was set where the water was 19 m. deep, but the net was fastened to eight weighted 11 m. lines, so that it floated just above the thermocline; a third was set on the bottom where the water was 8 to 9.2 m. deep; the fourth was set on the bottom at a depth of 3m. All of these nets were set at right angles to the shore line and were placed in a straight line from deep to shallow water. ‘They were left in the water three hours (9.45 to 10.30 a. m. to 12.45 to 1.30 p. m.). Nothing but perch was taken in the nets, and the catches were as fol- lows: On bottom at 19.2 m., 0; at a depth of 8 m. above bottom 19 m. deep, 0; on bottom at 8 to 9.2 m., 118 (49 alive, 69 dead; males, 42 dead, 17 alive; females, 27 dead, HABITS OF YELLOW PERCH. 331 32 alive); on bottom at 3 m., 5 (all alive; 3 males, 2 females). This experiment indi- cates that perch are bottom fishes at all seasons. The observations of Hankinson (1908), tii i (1915), and Meek (1916), in other lakes make it apparent that this condition is general. Meek (1916) states that perch are more sluggish in winter. ‘The gill-net and line catches for both lakes support his view (Tables 2 to 5 and 27). ‘The catches in Lake Wingra indicate, however, that cool or stormy weather does not interfere with feeding if food is available. On windy days, when the gill nets caught little, the usual numbers of fish were captured from a drifting boat on hooks.. The fishes in this shallow lake were apparently ready to eat if food was present but were unable or unwilling to move about much during storms. On several occasions schools of perch were observed at the surface. ‘This occurred once at 10 p. m. on Lake Wingra and was observed several times from 5 to 7.30 a. m. in Lake Mendota during the warmer months. As such schools were usually observed during early morning hours, it was thought that there might be a daily migration which would take the perch into shallow water at night and into deep water during the day. Such a migration could not, however, be very extensive, because perch caught at depths Fic. 35.—Positions of gill nets set to determine the comparative numbers of perch at different depths. of more than 10 m. were apparently unable to make rapid modifications in their swim bladders so as to become adjusted to surface conditions. When kept in shallow aquaria such deep-water perch, though apparently in good condition, often floated belly up at the surface for two or three days. It was possible, however, that there might be rhythmical migrations, a few meters in extent, with the changes accompanying day and night. Gill nets were accordingly set to discover if such were the case. They were arranged to catch fish at the surface and on the bottom, so as to give opportunity for comparing the numbers present in two or more situations, and were examined at the end of 4-hour periods for 24 hours. On August 12, 1916, three nets were set in Lake Mendota (fig. 35). One floated at the surface; another was on the bottom directly beneath it at a depth of 7.5 m. (just above the thermocline); another was inshore from the other two and on the bottom at 2.9m. The catches for this and two other similar experiments are shown in Tables 28 to 30. It will be noted— 1. That there were never many perch caught in shallow water near shore. 2. That in the bottom net just above the thermocline the catches in the early morning hours’ (1 to 3 a. m.) were usually the smallest. For the three experiments the average catches were as follows: From 12 m. to 4 p. m., 43; 4 p.m. to 8p. m., 56; 8 p.m. to12p.m.,37; 12p.m.to4a.m.,12; 4a.m.to8a.m., 41; 8a. m. to 12 m., qt. 110307°—21——22 332 BULLETIN OF THE BUREAU OF FISHERIES. 3. That the only time when perch were caught in the surface net was at 5 a. m. These results indicate that perch migrate from the region of the thermocline toward the surface during the night, but the number of observations is small and should be extended. In the experiment summarized in Table 30, where the deep net was exactly on the thermocline, not a single perch was caught during the early morning hours. All catches in Tables 28 and 30 marked with an e were taken ashore; all others were thrown back as soon as they were removed from the net. It will be noted that, when the fish were not put back, the next catch was not appreciably smaller. The first catch in each experiment should have been larger, if all other conditions were the same, for the nets remained in the water exactly 4 hours. Pulling a net and the removal of the fish occupied from 2 to 25 minutes, which would make the periods of time for the various catches after the first somewhat less than 4 hours. The fact that as many fish were caught during the next 4 hours, when an entire catch was removed from that region of the lake, as when they were put back indicates that, though perch keep to a particular depth, which varies somewhat with the time of day, they do not remain in one locality, but continually swim along the shore.* One other aspect of the migratory activities of perch remains to be considered. This is their habit of swimming in schools. Meek (1916) states that soon after hatching certain species of marine fishes form schools which retain their unity for several years. He also says that schools of fresh-water fishes are much more likely to mix; fishes of different ages, and even of different species, may keep together. Schools of perch have been observed at various times in shallow water in the two lakes under consideration in this paper. These usually consisted of fish of about equal size—large, medium, or small. They have been seen alongshore, among aquatic plants, and in the open lake both at the surface and at a depth of 3 or 4m. For example, during the latter part of August, 1916, a school of about a thousand young perch remained alongshore in one locality, just north of the University of Wisconsin, for more than a week. Hankinson (1908, 1916) and Reighard (1915) report similar schools of young perch in Michigan lakes. Catches in gill nets also indicate that perch swim in schools when in deep water. A net set in one spot and examined at intervals might catch nothing for several hours and then be filled in a few minutes (Table 30). A similar thing often happened when fishing with a hook and line. Furthermore, a 60-foot gill net might have 50 or 75 perch in one end and not a single individual in the other. All these observations signify that perch swim in schools throughout life. ENEMIES AND PARASITES. PREDATORY ENEMIES. In the Wisconsin lakes perch pay the penalty for exceeding other fishes in abundance by being preyed upon by a number of predacious animals. Among the fishes the pickerel (Esox lucius) appears to be the species which most commonly feeds upon perch. During the year 1916 the following records were secured from Lake Mendota: @ During the summer of 1917 additional evidence was secured which supports this view. Nine hundred and sixty-six perch were caught in gill nets at three stations in Lake Mendota. An aluminum tag was fastened to the dorsal fin of each, and they were then returned to the Jake. Although fishing with nets was continued for a total of 33 days at the three places, only one of the tagged fishes was caught a second time. HABITS OF YELLOW PERCH. 333 Perch eaten, Length of Date, tl picketel ) millimeters). Size Number. (millimeters), - a B= HHO ‘oO a The remains of perch have also been found in largemouth black bass (Micropterus salmoides) caught in Lake Mendota. Dogfishes (Amia calva) were often caught in gill nets in shallow water, and in many cases they were near perch which had been previously captured in the net. Such occurrences indicate that dogfishes may feed upon perch, but the authors have never found them in the alimentary canal. The gar (Lepisosteus osseus), doubtless, also feeds on young perch. Hankinson (1908) found pickerel feeding on perch and also mentions an 8-inch perch as occurring in the wall-eyed pike (Stizos- tedion vitreum). Forbes and Richardson (1908) state that 75 per cent of the food of the lota (Lota maculosa) is made up of perch. Reighard (1915) reports perch feeding on each other. Besides finny enemies, perch are probably often beset by other predators; for instance, water snakes, garter snakes, and bullfrogs may catch the young alongshore. Turtles often eat perch caught in nets, and probably feed upon them when they have a chance under natural conditions. A. R. Cahn has furnished observations on birds which eat perch. In Wisconsin he has found the following feeding on perch: Herring gull, common tern, black tern, Ameri- can merganser, red-breasted merganser, great blue heron, green heron, black-crowned night heron, loon, horned grebe. He states that the following also probably eat perch: Double-crested cormorant, white pelican, other species of gulls and grebes, and the bald eagle. Fisher (1893, p. 32) reports the fishhawk as feeding on perch; Eaton (1910, p. 137) mentions the kingfisher. The senior writer on June 10, 1916, saw a crow pick a crappie (Pomoxis sparoides) from the surface of Lake Wingra. Though the fish in this instance struggled actively and finally escaped, the crow may at times be more successful in its aquatic forays and capture fishes from the water. Probably such carnivorous mammals as the otter and mink at times capture perch. Among the predatory animals mentioned the only ones which commonly follow the perch into deep water are the pickerel (Reighard, 1915) and the lota. ‘The latter does not occur in either of the lakes discussed in this paper but is important in the Great Lakes and some other smaller bodies of water. The majority of the perch in Lake Mendota are therefore free from attack by predacious enemies during most of the year, except for an occasional pickerel. PARASITES. While the routine weekly examinations of perch were made primarily for the purpose of ascertaining the nature of the food, after March, 1915, a careful record was kept of the presence of parasites. This record is doubtless incomplete; the numbers are too small rather than too large. For example, many of the intestinal distomes were doubtless overlooked, because the food was stripped from the intestines, and they may 334 BULLETIN OF THE BUREAU, OF FISHERIES. have remained attached to its wall. The commonest intestinal distome, Bunodera nodulosa, lives in the bile ducts and gall bladder during early stages, but no regular examinations were made to discover its presence at seasons when it was not in the intestines. Every parasite observed was not identified as to species, but practically all, if not all, will fall in the list which follows. No routine record was kept of the occurrence of the skin parasite, Diplostomulum cuticola. The results of the routine examinations for parasites are summarized in Tables 31 and 32. Nematodes were never present as intestinal parasites during December; in Lake Mendota they were most abundant in summer; in Lake Wingra, from March to May and from August to November. In Lake Mendota no trematodes were found in the intestines during September, October, and November; and in Lake Wingra none were found at any season. ‘The cysts of larval proteocephalid tapeworms were prevalent in the liver, and often in the peritoneum elsewhere, during every month of the year. Larval proteocephalids were most abundant in the intestine from March to May in Mendota but were irregularly distributed through the year in Wingra. Acanthocephalans were most abundant in spring in both lakes. Leeches and adult tapeworms were uncommon and irregular in their occurrence. The most striking difference in regard to parasites between the perch of Lake Mendota and those of Lake Wingra is in the complete absence of intestinal trematodes from the latter. This may be due to the absence of a proper intermediate host in Lake Wingra. The following list includes all the parasites known to occur in the perch from Wisconsin lakes: PROTOZOA. Henneguya wisconsinensis Mavor and Strasser.—This myxosporidian was first described from speci- mens taken from the urinary bladder of a male perch caught in Lake Mendota and examined on April r5, 191s. During the present investigations no examinations for this parasite have been made. CESTOIDEA. Proteocephalus pearsei La Rue.—Specimens of larval cestodes, cestode larval cysts, and adult tape- worms were sent to Dr. G. R. La Rue, of the University of Michigan, who was kind enough to describe them (1919). One of the larval cysts was found in the body muscles on October 13, 1916. TREMATODA. Bunodera luciopercae (O. F. Miiller).—This fluke was common in the intestines, particularly in the ceca, in perch collected from Lake Mendota but was absent from those collected from Lake Wingra. It has previously been reported in the perch from this country by Stafford (1904) at Montreal, Canada, and by Marshall and Gilbert (1905) from the lakes near Madison, Wis. Clinostomum marginatum (Rudolphi).—This trematode was observed twice in the perch from Lake Mendota. On September 25, 1915, a cyst containing a nearly mature specimen was found beneath the skin in the flesh at the base of the tail. On January 10, 1917, the gills of 20 perch, which had been caught at a depth of 17 m., were examined and one small larval cyst was discovered, embedded in a gill filament. These isolated observations, of course, give no idea of the prevalence of this parasite in Wisconsin. Diplostomulum sp.—This skin parasite was observed now and then in the lakes near Madison and was always more abundant in the young fish than in adults. It was very prevalent in the perch from Oconomowoc Lake. An idea of the difference in infection in perch from two Wisconsin lakes may be gained from the following statistics: Fourteen perch, collected from Lake Mendota, near the base of Picnic Point, August 24, 1916 (length—maximum, 69; minimum, 52; average, 61 mm.), were infected to the degree shown by the following ‘‘number of infected individuals—total number of parasites—average”’ figures: Tail, 5-5-3.5; fins, 0; head, 8-16-1.1; ventral region, 9-18-1.3; dorsal region, 14-15-1; whole body, 1455-4. HABITS OF YELLOW PERCH. 335 Thirty-nine individuals collected on July 17 and August 8, 1916, in Oconomowoe Lake (length— maximum, 98; minimum, 35; average, 55.1 mm.),showed: Tail, 25-53-1.3; fins, 24—79-2; head, 26-98-2.5; ventral region, 31-132-3.3; dorsal region, 25-123-3. 1; whole body, 38-490-12.9. (?) Allocreadium, isoporum Looss,—One specimen, which is apparently referable to this species, was found in the intestine of a perch collected in Oconomowoc Lake, August 14, 1916. Crepidostomum cornutum (Osborn).—Eight specimens were found in a perch caught in Lake Men- dota at a depth of 18 m., January ro, ror8. ACANTHOCEPHALA. Neechinorhynchus cylindratus (Van Cleave).—This was the common acanthocephalan found in perch. Sometimes it occurred in great numbers. In one instance a couple of hundred were found in the intestine of a single perch. Echinorhynchus thecatus Linton.—One specimen, which is apparently referable to this species, was saved from a perch 60 mm. long caught on August 24, 1916, in a minnow seine east of Picnic Point. NEMATODA. Dacnitoides cotylophora Ward.—Specimens of the nematodes from perch intestines were probably of this species. HIRUDINEA. Piscicola punctata (Verrill).—This was the species of leech usually found on the perch in the lakes investigated. Placobdella parasitica (Say),—One individual of this species was found attached to a perch caught in Lake Wingra October 28, 1916. Our thanks are due to Prof. J. P. Moore, who identified it. INSECTA. Psephenus sp.—On July 22, 1916, a perch, caught at a depth of 15 m .in Lake Mendota, had a- “water penny’’ attached to its body just behind the right pectoral fin. This beetle larva must, there fore, be recorded as an accidental commensal or parasite. GENERAL DISCUSSION AND CONCLUSIONS. The investigations on the perch in the two lakes selected for study have been described, and it is now time to return to the problems it was hoped they would solve: (1) To account for the abundance of perch compared to other species of fish; (2) to determine why perch have a particular maximum size in certain lakes and why they are larger in some lakes than in others; and (3) what effect stagnation has on the activities of fishes. The perch appears to be more abundant than other species of fish because it is versatile and not too specialized. Though it has certain specificities of behavior, such as the habit of usually feeding on or near the bottom, it is able, more than any other fish with which it is associated, to invade all habitats. It may feed on the enormous quantities of plankton in the pelagic regions; it is at home among aquatic vegetation; and it may grub out the animals embedded in such great numbers in the soft bottom mud or even largely subsist for a time on the mud itself. Its chief advantage over the common shore’ fishes is in its ability to forsake the shore, with its stores of food dependent chiefly on the aquatic vegetation, and invade the depths of the lakes, where the chief source of food is the soft sedimentary bottom deposits rich in organic constituents. The perch has rivals in each of the habitats where it seeks food, but it is an able competitor of them all. In shallow waters it may capture mollusca as well as the pumpkinseed, littoral plankton as well as the silversides or bream, insects and their 336 BULLETIN OF THE BUREAU OF FISHERIES. larve as well as the bass, crayfishes as well as the dogfish, small minnows as well as the gar. In the open lake the perch’s chief competitors for food are the cisco and the white bass, but neither of these fishes excels it in ability to strain plankton from the water. In the deeper regions of lakes the perch must contend with the vegetarian and bottom-feeding sucker, cottid, and carp, and with the predacious pickerel and lota. The sucker, cottid, and carp are real rivals when it comes to bottom feeding, for they are especially able to take advantage of the nourishment in the bottom mud.? They are also better protected, by reason of their size, from the attacks of the predacious deep-water fishes; but their large size, on the other hand, limits their numbers, and they can never compare with the perch in this respect. These bottom feeders are limited, however, in times of scarcity or when they are driven into shallow water by stagnant conditions in the depths. They can not then feed as well as the perch in pelagic or littoral regions. Perch are, then, more abundant in lakes than other kinds of fishes because they are of intermediate size and because they are better able to secure food from all ayail- able habitats and at all seasons of the year. There are probably a number of factors which cause perch to attain a certain characteristic maximum size in different lakes. This is a phenomenon which is not confined to perch alone but has been noted in other fishes. It is apparent, for example, in the ciscoes in various Wisconsin lakes, and has been observed in other localities in various parts of the earth. Petersen and Jensen (rg11) state that the plaice in a certain estuary ceased to grow for two-thirds of a year, whereas some which were transplanted quadrupled in size during the same period of time. They believe that the discrepancy in this instance was due to differences in food. The present authors believe that their comparison of the habits of perch and their conditions of life in Lake Wingra and in Lake Mendota have shed some light on the causes for such contrasts, and they feel that they can, at least in part, give specific reasons why the perch are smaller in the former lake. The shallowness of Lake Wingra is probably the chief cause for the small size of its perch. ‘The limitation of perch to a stratum of water 3 m. in thickness, between the air above and soft muddy bottom below, causes many unfavorable conditions. Winds stir up the whole body of water; thus movement and feeding are often made difficult or impossible. Knauthe (1907) has made the generalization that in two ponds of equal capacity in other ways the quieter one will be the more productive for rearing fish. On account of the shaliowness of the water in Lake Wingra there are wider and more rapid variations in temperature. The water is all warm in summer; there is no possibility of retreat into cool, quiet depths; and consequently the perch in this lake pass through a period in late summer when little food is eaten (Table 22). In winter, the perch in Lake Wingra move about very little and hence feed less than those in Lake Mendota (Tables 2 to 5, 22, and 27; figs. 3 to 5). Though oxygen was always present in quantities sufficient for respiration, many of the fish caught in gill nets in Lake Wingra died during the warmer months, when the water was murky with alge and other organic or sedimentary products. The perch in Lake Wingra, on account of the earlier warming of the water, breed before those in Lake Mendota, when the season is less advanced and food is less aT he importance of this deposit as a source of food has been pointed out in a masterly wav by Petersen and Jensen (1911). HABITS OF YELLOW PERCH. 337 abundant. They also mature the gonads earlier in the autumn, in part during the hottest weather when food is readily available but when feeding conditions are unfavorable. Perch hatching in Lake Wingra have less desirable conditions for feeding during their growth period. There are two conditions which appear to be more favorable to the perch in Lake Wingra. One of these is the fact that there is abundant oxygen for respiration at all depths during the summer when feeding is active. The other is the entire absence of trematode intestinal parasites. These two factors, however, appear to be of little importance compared with those cited in the preceding paragraphs, which are more favorable in Lake Mendota. The differences between the perch in the two lakes in regard to the constituent elements in the food are probably not important in determining maximum size. As has been stated, the perch in Wingra eat more of insect larve and less of Entomostraca than those in Mendota; but there is no reason to believe that such differences would account for the discrepancies in size. The chief generalization to be made from the comparisons between Lake Wingra and Lake Mendota is that, at least in temperate regions, a deep lake is a far better habitat for most fishes than a shallow one and will usually be more productive. There is no doubt that some fishes, such as the crappie, are peculiarly adapted to shallow- water habitats, as shown by the senior author in a report (1919) compiled from studies, extending through an entire year, of the abundant crappies in Lake Wingra. But though there are such special cases and though more extensive studies in different types of lakes will doubtless bring new facts to light, the authors believe their first statement will, in general, hold good. Of course, a deep lake without suitable breeding grounds and with a scanty fauna would have few fishes, but even under such circum- stances it would excel a shallow lake with similar characteristics. One other problem remains for solution, and, though the results presented in this paper do not solve it, they may help to do so. This is the determination of the factors controlling the productiveness of lakes of various types and sizes. An understanding of this may in time give man the power to control and increase production. An attempt has been made to gain some idea of the productiveness of Lake Mendota in terms of the total number of perch caught from its waters per year. Daily counts were made of the number of fishermen on the eastern half of the lake from January 8 to February 27, 1917, at 10 a. m. and 3 p. m. At intervals trips were made around the lake to ascertain how long each man had been fishing and the number of perch caught. The number of fishermen averaged 19.1 in the morning and 31 in the after- noon. ‘Their catch per hour averaged 23.6. Estimating that each man counted fished three ‘hours, and that one-fourth of the fishermen (not counted) were on the west end of the lake, the total average catch per day for all fishermen would be 2,358. There is more fishing for perch in winter than in summer. Probably the number of perch caught per day when there is no ice is about one-fourth that during the winter. This means that 2,358 perch are caught per day for four months and 589 per day for eight months. From such speculation it may be estimated that 424,540 perch are caught each year from Lake Mendota. Judging by the other fishes taken in gill nets and by our general knowledge of conditions, it is estimated that the total annual catch of other species by fishermen is about as follows: Pickerel (Esox lucius), 2,208; white bass (Roccus chrysops), 615; rock bass (Ambloplites rupestris), 613; silver bass or crappie 338 BULLETIN OF THE BUREAU OF FISHERIES. (Pomoxis sparoides), 183; largemouth black bass (Micropterus salmoides), 305; pumpkin- seed (Eupomotis gibbosus), 428; bluegill (Lepomis incisor), 1,238. This gives a rough approximation of the number of food fishes caught in Lake Mendota each year and may serve as a standard for lakes of similar size, depth, and situation. The old fishermen claim that many more fish were caught 15 years ago and state that a single man sometimes secured over 800 perch in a day. At present the usual catch of a professional fisherman, fishing through the ice with a line and two hooks, is from 200 to 400 per day. Lake Wingra not only has smaller perch, but fewer of them. ‘This is clear from the eatch per hour in gill nets (Tables 2 to 5; figs. 3 to 5). The reasons for lesser size have already been discussed, and apparently the same reasons set a smaller limit to num- bers. The differences between the sizes and numbers of perch in the two lakes are due to variations which interfere with growth and allow fewer individuals to survive in Lake Wingra. SUMMARY. 1. The habits of perch in a small, shallow, and muddy lake were compared with those of perch in a neighboring large, deep, and clean lake. Perch were the most abun- dant fishes in both, but, in proportion to the size of the lake, there were more in the larger lake. 2. The perch is a versatile feeder but usually gets its food on or near the bottom. The percentage by volume of the foods eaten by 1,147 adults was as follows: Chiro- nomid larve, 25.2; cladocerans, 22.1; Corethra larve, 6.4; silt and bottom débris, 6; chironomid pupa, 5.9; fish, 5.2; amphipods, 3.6; Sialis larve, 3.4; caddisfly larve, 2.1; oligochetes, 1.5; crayfishes, 1.5; odonate nymphs, 1.4; clams, 1.2; alge, 1.2; snails, 1.1; ephemerid nymphs, 0.9; calcium carbonate crystals, 0.5; leeches, 0.4; hemipterous adults, 0.3; mites, 0.3; chironomid adults, 0.2; Corethra adults, 0.2; Corethra pupa, 0.2; copepods, 0.1; ostracods, 0.09. 3. There are more or less marked seasonal variations in all constituents of the perch’s food. In general, foods are eaten in proportion to their abundance and avail- ability; but this is not always the case. 4. An adult perch eats about 7 per cent of its own weight each day. Digestion is three times more rapid in summer than in winter. 5. Perch do not take any abundant food but select certain things. There are daily and seasonal variations. Individuals feeding in shallow water eat a greater variety than those from greater depths. Perch contain food which is available at the depths where they are caught, which indicates that extensive vertical migrations are infrequent. 6. Food varies with age. During youth there is a change from Cyclops and other entomostracans to Hyalella and insect larva. At the end of the first summer the food of young perch is much like that of adults. 7. As judged by the rate of increase in young perch when fed on a single food the following varieties rank in the order given, the best being first: Earthworms, ento- mostracans, chironomid larva, amphipods, fish, small amounts of various foods, liver and flour, adult insects. 8. In the small lake investigated insects were the most important constituent of the food. In the larger lakes insects and entomostracans were equally important. HABITS OF YELLOW PERCH. 339 g. Compared with the crappie, the perch eats a greater variety and shows other specificities of behavior. 10. Though perch are able to recognize the proportions of oxygen and carbon dioxide in water, they enter regions where conditions are unfavorable for respiration and may remain in oxygen-free water for as much as two hours without dying. Whenin water without oxygen perch use part of the oxygen in the swim bladder. 11. Perch may become sexually mature in two years. In the smaller lake inves- tigated they generally become mature when of much smaller size than do those in the larger lake. 12. During the spawning season the males come into shallow water and remain for some time. The females remain on the spawning grounds only long enough to breed. 13. Except during the spawning season and when the deeper water is stagnant, most of the perch in the large lake remain in deep water through the year. In the smaller lake similar migrations take place. 14. There appears to be an upward migration at night. 15. Perch swim more or less in schools throughout the year and apparently do not remain in one locality but move along the shore. 16. Perch have many predacious enemies. The pickerel and lota are important. 17. Perch are very generally infected with parasites. Those in the two lakes investigated contained cestodes and cestode larve (one or more species), trematodes (5s), acanthocephalans (2), nematodes (1). Leeches and an insect larva were found on the outside of the body. 18. Perch are more abundant in inland lakes than other species because they are more versatile. 19. Large inland lakes will generally contain more fishes per unit of volume than those of smaller size. 20. Judging by the data presented in this paper the reason why the fishes in certain inland lakes attain a rather small maximum size is because there are various adverse conditions which prevent growth. In the present instance food does not appear to be as important as other factors, such as shallowness, exposure to wind, etc. TABLES. TaBLE 1.—CoOMPARISON OF LAKE MENDOTA AND LAKE WinGRA, EAcH 258.8 M. (849 FEET) ABOVE SEA LEVEL. Length. Breadth. Area, Maximum depth.| Mean depth. Shore line. Lake. : : Square A Kilo- . Kilo- . - Square Kilo- . meters,| Miles. | meters, | Miles. — miles, | Meters.| Feet. | Meters.) Feet. meters, | Miles. Mendota..... 95 5-8 4 45 39-4 15-2 25-6 84 12.1 39-6 32-4 20.1 Wingra 2.6 1-6 1-4 8 2.17 +79 4-25 14 1.6 5-3 73 45 340 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 2.—ToTaL AND ComMPARATIVE NUMBER oF PercH Caucut in Gu Nets SET at VARIOUS Depts 1n Lake MENDoTA, JULY 2 To DEC, 14, 1915, WITH NoTES ON THE AMOUNTS OF OxYGEN AND CARBON Di0xIDE PRESENT. Cubic centimeters of gases per liter. Depth net set (meters). Date. oO. CO. Number of perch caught, Average per hour at— Total. o-5 m. 5-10 m. Io~IsM.| 15-24 m, ~7 |. 31 19-0 54 6.0}. I2 3-5 |- ° 18.5 |. 94 br8.5 |. 17 INOW urges adn ciawe heen b 13.0 |. 20 13-0]. 94 40]. ° 3-0}. ° 14-6 |. 23 3-5 |. 7 3-5]. I 16.3 |. 41 2r.0}. 32 3-5]. 2 12-5 |. 47 35 2 ©909700000000000000000008aNn @ All nets used were 1-inch bar mesh, measuring 3 by 75 feet, except those marked >, which were 34-inch bar mesh, measuring 4 by sofeet. For some reason the latter caught only one-fifth as many perch as the former, as is demonstrated by the catches on Noy. 6 and 13. » See note under a. HABITS OF YELLOW PERCH. 341 TABLE 3.—PrERcH CauGHT PER Hour AND PER Day 1n Gu Nets, Lake MenpotA, 1916. Cubic centimeters of gases per liter. Gill nets, Number of perch caught. Depth set mir (meters). Avera; A . A Hours in Pees $53 CO:. oO. Time set. | Time pulled. Reuter: Total. per per 24 P our, hours, DESH a stein ale ciktched wictatarct aiainle [atAS aMteieh rity s vf nina iciple ste tia te sila ovate tinge Estee 6 19 a4 18 O7 14-5 24 18 7 BOUL Neds ciainrw ellel Matahh fleidielnie tetetasolelotabs oft, «a's | wiafalera@ cate Ren acinty 24 36 5 17-3 5 pm. .m. 14 8 “5 5 . 5 p.m. .m. 14 6 +4 a18.2 |. 4-45 p.m. .™m, 13-7 14 1.0 . 18.2 4-45 p.m. .m. 14 3 2 5 a3 -1.8 4-45 D-™m. . .m. 14 163 11.6 “ 3 ~1-8]. 4-45 D-™. . . mm. 14 380 27-90 a AST. 90... d.cccbeccess 14-5 |. 6.50 a, m. 10.05 a. Mm. 3-2 a «15 3-6 12-5 |. 7-05 a. Mm. 10.14 a. ™. 31 6 +3 7-2 8.5 |. 7-16a. m. 1o.20a.™m. 3 8 “5 12.0 5 7.25 a.m. 10.32a.m, 3-2 Ir 3-6 86. 4 SF ee A aah 15 1.53 D-™. 4-53 D-™. 3 12 4-0 96.0 wr 2,08 p.m. 5-08 p.m, 3 17 5I 134-4 4:77 45 2-18 p.m. 5-20 p.m, 3 23 7-6 182.4 5 4-1 2.22 p-m, 5-16 p.m. 3-1 18 6.0 144-0 MOY: 65 5.chois estes vase 6.38 a.m. 9-38 a.m. 3 7 23 55-2 6.51 a, Mm. 9-51 a.m. 3 7 2-3 55.2 7-00a,m. 8.00 a, m. I 8 8 193 714 a.m. g-10a.™m, 2 I “5 12.0 May 12...... on enpee 5:12 a.m. 8.12 a.m, 3 14 4.6 Ilo. 4 §.20 a.m. 8.21 a.m. 3 32 10.6 254-4 5-31a,m, 8.36a.m, 3 9 30 72.0 5-41 a.m. 8.42 a.m. 3 53 17-6 422.4 May 17...... “pacdeor 12.01 a, Mm. 2.00 a. M1. a 47 23-5 364.0 12.13 a.m, 2.13 a.m, 2 6 3:0 72.0 bie Oe Re eee 8 19a,m.| r1.19a.™m. 3 8r 27.0 648.0 8.33 a. Mm. 11-35 a.m, 3 75 25.0 600. 0 8.43 a.m. 11.44 a.m, 3 53 17-6 422.4 8.57 a.m. 11.57 a.m, 3 I 3 73 9-01 a.m. 12.05 a. M. 3 ° ° ° DS ACESS SRL Sprpsce| 10.42 a, ™m. 1.53 P-™. 3-1 30 96 230.4 10-5 a.m, 2.00 p-™. 31 145 45-8 1,099.2 11-07 a.m. 2.05 p.m. o 26 8.6 206.2 II.20a.m, 2-11 p-™m. 2.8 9 3:2 76.8 July 6...... aeeaeamcls I-00 a.m. 1.46 p.m. 2-7 28 10.1 242-4 11-10a, mM. 1.56p. Mm. 2.7 126 40.5 972 II. 20a. mM. 2.03 P- ml. 2-7 49 17-8 427.2 1I.30a. mM. 2-10 p-™, 2.6 4 1-4 33-6 July 14...... Loeadnad 8.35a.M.| 10.35 a. 1. 2-0 Ir oS°5 132-0 9-08 a.m, 11-08 a. Mm, 2 77 38-5 924.0 9-48 a.m. 11.48 a.m. 2 62 31-0 744-0 10.14a.m. 12-14 p-™,. 2 6 2.0 48.0 July 22 ba Mbaieo wet 7.27 a.m. 9-33 a.m. 2.t I “5 12.0 8.14 a.m. 10.14 a. Mm. 2-0 a9 14-5 348.0 8.44 a.m. 10.44 a.m. 2.0 1a 6.0 144.0 9-10a.m. Irl.10a. mm. 2 2 1-0 24-0 ANG. 7, 00 che Sedacsesss 8-45a.™. | 10.45 a.m. a ° ° 0.0 9-15 a.m, 11-18 a.m. 2 ° ° 0.0 9-43 a.m. Il. 44 a.m. a ° ° ° 9-30a. m. 11-35 a.m. a 15 TS 180.0 II. 53 a.m. 1.53 D-m. a ° ° ° PH Bi chive Reece cdl® | Be SaiZs Se vinnie Scheie od] sade ROA cic Mlcocadectlde cei b4 45 11.2 268.8 ARORA «6 ichandaatlen ne b4 122 30-5 732 AUErC. oi FFs. 00264 4 9-45 a.m, 12-45 p.m. 3 ° ° ° 10.00 a, Mm. I.02 p.m. 3 ° ° ° 1o.18 a.m. 1-138 p.m. 3 118 39:3 943-2 1o.30a,m, I.40p-m. 3 5 1.6 38.4 8.35 a.m. 12-33 P-™. 4 I +25 6 9-00 a, m. I.00 p. M1. 4:0 ° ° ° AUG 30s ode cPececues 9.128, m. 1.18 p.m. 41 Ir 27.0 648 12.33 p-™. 4-30D.m. 4:0 3 +7 16.8 5-00 p.m, 9.10 p.m. 4.1 ° ° ° 5-18 p.m, 9-18 p.m. 4-0 71 17-7 424-8 BMBF IGs ccs Pheccvcs 9-00 p.m. I.20a. m1. 4:3 I 2 4.8 9-10 p.m. 1.32a.m. 4:3 ° ° ° 9-18 p.m, I-43 a.m. 45 24 53 127-2 I.20a. Mm. 5-00 a. Mm. 3-6 I 3 7-2 1.32a,m. 5-10 a.m. 4-2 12 2.8 67.2 I. 43 a.m. 5-18 a.m. 3-5 72 20.5 492-0 5.00 a, Mm, 8.30 a. m. 3:5 I +2 4-8 5-104. m, 9-00 a.m. 3:8 ° ° ° 5-18 a. Mm 9-18 a. m. 4-0 76 19 456 All nets marked with this sign were 34-inch mesh, bar measure, 3 by 75 feet. All other nets were r-inch mesh, 3 by 60 feet. o ® Nets set in Catfish Bay. 342 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 3.—PERcH CAUGHT PER Hour AND PER Day In Git, Nets, LAKE MENDOTA, 1916—Continued. Date. AU 24 EBA ee AME 25. ..,.RK.... 4 AMP 26. 2 Oh. es AdsgRID7 25s SEK ewes Bngtgo. .. 5. bk... PTB lo Hits ASE AEE SEBEL oc. cg aces Septta....1 588... Si BO NGas Jodee Joanae NDE IOS. bo fgdes ess Depth set (meters). an 1 NOD n x 6. 5- oI IPO NDOOWO~ Ane O (xr ~t~3 53 00 Il. 2-11. 5 8 10 II. 5-11.6 II. 6-11. 7 (a) * Ae 2» UMA UD UDAUA UA UA uw 2. (2) Qe on 5 wm ic Cubic centimeters of gases per liter. CO. Gill nets, Time set. | Time pulled. 8.15 a.m. 9-40 a, mM. I.5 8.15 a.m. 9-30a, Mm. 1.2 8.15 a.m. 9-I5 a.m, I 8.10 a. m. 8.35 a, m. “5 8.00 a, Mm. Io. 20 a, m. 2-3 To. 40 a. Mm. 12.35 p-m. 2.0 8200 a, m. 8.50a, m. 8 8.55 a.m. 9-47 a.m. 1.8 9-53 a. ™. 10. 23 a, Mm. “75 12.30 p.m. I.30p-m. I-o I-35 p-m. 2-35 p-m. 1.0 2.40 p.m. 3-30p.-m. -8 8.15 a.m. 9-00 a. m. “95 8.25 a.m. 9-15 a, m1. 8 8.00 a. m. 9.28 a.m. I.5 8.10 a.m. 9-23 a.m. I.2 9-35 a. Mm. I.2 Io. 30 a, Mm. 8 Io. 33 a. M. 8 Io.36a.m, -8 12.00 mM. 1-3 12-10 p.m. 2.00 p.m, 1-2 6.10 a.m. 7-35a.™. 1.4 6.10 a.m. 7-30a.m, 1-3 6.10a,m. 97-25a.m. I.2 6.15a, m. 7-35 a.m. 1.3 6.142. mM. 7-55a.m. 1.6 6.12 a.m. 8.04a.m. 1.8 6.23 a.m. 7-04 a.m 6.29 a.m. 7.08 a.m. 7-20a.m. 7-58 a.m. 6.22 a.m. 7-308. mM. 6.25 a.m. 7-25a.m 6.24 a, mM. 7.30a.m. 7-35 a.m. 8.44 a.m. 6.30a,. m. 7-45 a.m. 6.34.a.m. 7-38 a.m. 9-43 a.m I. 43 p-m. 9.49 a.m 1.52 p.m. Io.10a.m 2.04 p.m. I. 45 p-m. 5-37 p.m. 1.52p.m 5-48 p.m. 2.06 p.m 5-55 p-m. 5-43 D-™ 9-47 D- Mm. 5-48 p.m to. 08 p.m, 5-55 p.m to. 18 p. m1. 4-3 10.06 p.m 1.47a.m 3-6 Io. 10 p.m 1.44a,.m EAL 10. 20 p.m. 2.13a.m 3:9 1.45a.m 5-38 a. Mm. 3-9 2.05 a, mM 5-48 a.m 37 2-15a,m 6.03 a. mm. 3-8 5-41 a,m. 9.43 a.m 40 5.50a.m 9-55 a.m. 41 6.05 a.m. Io. 10a, M. 41 9-508. mM I. 43 p.m. 3-9 11,15 a.m 3-07 p.m, 3-9 II.25a.m. 3-25 p.m. 490 11.45 a.m 4.00 p.m. 43 3-25 p.m, 7-15 p.m. 3-16 3-26p.m. 7-33 p-m. 4-1 4.02 p.m. 7.42 p.m, 3.6 7.30pm. II-15 p.m. 3:75 7-35 p.m. II.20 p.m, 3-75 7-45 p.m. Il. 32 p.m. 3-75 II. 17 p.m. 3-06 a.m. 3.8 3-12a,m. 7.06a,.m 3-9 3-17. a.m. 7-20a.m 4.0 3- 29a, m. 7-37a.m 41 7-17 a.m. 11.15 a.m. 4.0 7.23 a.m. II. 23 a.m. 4.0 7-38 a.m. II. 45 a. Mm. 41 1.55pm. 3-25 p.m. 5 2.10 p.m. 3-40p.m. 1.5 2.19 p.m. 3-49 p.m. 1.5 2.27 p-™. 3-57 p-m. 1-5 Hours in Total Io I o CI 4 w& CoSc000b 0 OF OC On OOS NO BHO a & OrFOOHODOHODCOOOO0O0 » I Number of perch caught. Average | Average per hour. 12.6 137-6 wu see oo P- CO OA no CI a 800000000 mOHOHHO 4 un ey HH 2S mre » OPpOowWOOwODO0O s wo H w » e©o00000n00ON00 H I ono nn @ Surface. per 24 hours. 4 y Voss Bopoo000000 oo o i) S osoopon n 343 HABITS OF YELLOW PERCH: TABLE 4.—FISHES CAUGHT PER Hour 1n Gr Nets, Lake WINGRA, 1916.4 7T “STISOy ooooo ° C©CAO0DDDDDDDADDDDDDOODOODODDOOODDDODODOODOOOOOOOOOONDOOO0OOODOOOOOOOOOKNDOOOoOKMa OO -ngau sninieury man .=) . “BATES BIO 002000 0 000000000000000000000000'° “*c00000000000NDDO00000000000000 CO0HDONDOOOHKHOOOO *snsoq uv ” 00 -q13 snomodny 00000 0 0000000000000000'0x00000 0O00TD0D000000000000000000HOHHOD0000000HD00000H 000 Se 2° +0 ‘ordivo snotid sD 00000 0 0000000000 °0000000000000000000000000000000H000000000000 0 0000000006 n000 “seona] Hom Osho stmeqy 00000 0 00000%000000000000000000% “000000000000000000000000%000000000000000000000 *saprour OH mm 0 ” 4es snuajdoniyy | 99900 © 00000000 ©LKXOODOOFODDHDOO00D000000000D0000000000000000HHD GO00000000000H%0 “"Naraojop MS a) sniajdorstw]| 999090 0 0°000000000000000000000000000000000000000000000000000000000006000000000% “sniasso snajysostda’T 00000 0 000000000000000000000000000000000000000000000000000000000000000000000000 *sapror ” ) ” c) © -eds s{xowog 020000 0 04H000000H00000000000000000 00000nR0D0000HOO HONDDO000T00 0O000H 00000000000 .-) uv an is] omnn 2 oa “rOSIOUY sTm0da’T 00000 0 ONOH0000"'0 0O000000ND000 “OHODDDD00D0000D0000000000000000 HHODO' un OtO0000000 “ bel ao ho) “sno Xosy ©o "00H 0 0X 04400000000 °00000000000000000000000H000000000000000 c0000000000000000 Lad ” a a o p> a“ Ae) ” m oO a . “sTIaOSeABE BOIIT m° O90 © HOO "OMDDDD “DOOHDDDDACODDOMOOMNDDDOODDD0OSODORODDD000DOMO2D0000HORHHUNHADODOD0HO 8 Ms TT) " n rt mas ur smoyt t3ta0 " ASST HH EMM HANA” THK AMM HHH CRANK CR HR On ee” é heed uw ~ Ma Na NN 2 | -G@eqpm) Bo age ARS as ATS ATA AS LSS SAS LASS AD ASAE LASS LATS RRS RRS AER, = | yseutazis =| et i) *“(siajaur) eo » wnnn nin ne hmunnnn w win nw yes yydaq anna on ee i ee ee ee en er ee ee ee ee en a ee ee a on momen ono nono me memo m ener enone AG ; Go i ; ; § ; -fu20 saeisep) : “ere : y : “5, 2 S ‘ ainzeiadupy, ; eae = : - tad g os 5 ————————— ——— ————— o 5 5 : : a A ° =e x 5 a A : . A ener 2 0 x8 a ° we o e ko e a a a a a § < q< < > YB Cee Ope ae: OAR Ghats wieiselelals.s sing May 133..... May 20..... May 27 2 In all nets the mesh sizes were bar measure. The 1-inch mesh nets were 60 feet long; all others, 75 feet long. BULLETIN OF THE BUREAU OF FISHERIES. 344 TABLE 4.—FIsHES CAUGHT PER Hour In Git, Nets, LAKE WINGRA, 1916—Continued. Snsol 000000000000000000000000000000000000000 20000000000000000060000000000000000000 -ngeu snimary va BATeO BIUTY 229000 00H000000000000000000000000000000 22000000000000000000000000000000000000 “snsoq eo ” nw ist ~ ao o -q13_ sromodnsy 9000 2907000H0000000000000000000" "oH mo00 0909099000000 °0H900000000'000' 0000000000 bel oydiv snd | ©20000000020000000000000000000000000000 229000000 00000000000000000000000000000 “SEON3] 8 sh stureiqy ©9%DO0000000000000000000000000000000000 ©020200000000000000000000000000000000000 iseprous ‘ a jes sniajdomiyy | 29929000000000000000000'000000 000000000 ©000202D0000000NDDD0000000000000000000000 “naramojop o] =" te we a My sNiajdortyy | SC°X 00000000000 NDTD00000000000000000000000 9 *0000°000000°0000000" 00000 "0000000000 “sTlasso 9 5 Cc snajsostdoy 290000000000000000000000000'000'o0000000 20000" 00000000000000000000000000000000 = “sapror + 00 + mo 1° a ° -eds sIxomog 9©0°00°00000000000000000000* ‘c0000H000 290000000" 000d0000000000000000000000~° T+ oO mo ~ Jostony staroda’T 02+000004H000000000'0000000000"'0 0000000 ©0000000000000000000000000000000000000 + “snfony Koss 9200200000000000000000000000000000000000 9000000000000 c00000000000000000000000 = on io a aa an a 7 bel + ” o oO nw nowt Mw m oO StooseARy voIIg 0 0mmae 5000000000000 ad00KH0c00000000 000 Oo OmMAOO™ MOON OMD000HO*OH “0H “0 *000H0‘0 “JOJeEM a um Lalial MMMM runMnAa ArmMoOoonnnontd ner nn nrnanr i co ODAMNNH wormovonom So um Conn A HOnN Ur sinoxy CH CHAR AHA “HHA dN HM MHRA AHHH “AO Cm mmm dnt ee egeeHHt nana stemsetdannnmasee . L’ tal we a 4 1 St Nt Nt Nt a] SENSE NS NOE tN Nt Ne Nt Nt a] t so Nt Nt a) oN\t NS Nr ov wNt wt o a) Nt a] uv . - » ro) ® » Pir) » B | cGsoupemy | SS SR SS TT ee LS RW ee RL = | Yseurazig ; Jes Wydaq SHAH HHH RAR HR RK KKH Ka mam mmdaddddd mma MHA MHHAADHH AH AMAA MMA HH AK ARA HAHAHA ‘(epeis va) : + : co: 5 a » a) ) a =r & Hue seeidop) 6 : ° o > 4 6: : : oy mS ® + ¢ mo: ainjzelsd may a : 7) a : Ct a: : : " i ” m i « : 3 ee i i; — STEN eee A ; ri : : = : = ; : : : 4 : : : jon d . ‘ 5 : : ; , = : + 2 ° : . Ae 70 g Py GE} “8 2 2 “oR 8 (ah Soap emt Hoos q g > b> Cr) P) ] by eae 3 ; ; coe | 5 ih ose} Sood 3 3 ste days at a i eae Rea tou a 4 a 4 qo 4 4 < 4 2 hb & i Sis fe) ore) 345 HABITS OF YELLOW PERCH. TABLE 4.—FISHES CAUGHT PER Hour IN GILL Nets, LAKE WINGRA, 1916—Continued. ~0sh19 s1mIVIqy “snsoy 00000 -ngeu sninieury wap wy | 00000 eo00000000000DDVDADADADCCOAOA0OADAOAGGCA0AG000 ©o0000000000000000000000000000030 = *snsoq ° -qi3 snomodny | ° °°9000000000000000000000000000000000 $3 6 6S ordi sna | *“*900000°0*"0000000000000000000000000 = “SPono] ° eo0o000000000000000000000C0AC0CCC00C00000 *“sapiour “Jes stie}doi1y 0, 0'0 0.0 “narmrojop SNIIPGOIIIW DEO LOOKS *snasso snazsosidaT “saplor -eds = stxomlog PEA) *lostour srmmoda’T +02 ecooo0o0o000000000000000000000000000 eo0o0oc0oee0eecoeoeaeceeoeoeoeoeee0aeaeaeeeaeoAaeaeaea 00080 a) = = ° ° = snpny xosey 220000 000000000*000000000000000 “0000 8 HH bel %B mo -) ts Lal ”m . w o suaosaALg Vlog “90a °00°00'00M'00*00%'00"°00"°00'00'0'0 *Ig}eM owe Murnane an tun mun t+ | oO “ ae H HK wadad BaARaAGOR Ur sSMOH | SQrPPLamessaan BAA MSS SSeS SSS TANS ai. REx ¥ coh a ESE ES a (Sea ANED ia haces ss SS eae end as va eleg veh nies Lil cde Gas SoetdrenT eel PUKE Laas = «| Wseur ezis 1) “no vt wun wun GCN TE 1429 | eat ede) : Pr heh me heart yes WIdeq mmm HAMA MMA MMH AMM AMAR AMMO MMMM NOH He ‘(apeia ” wn ” “ o a) -1yU99 saaizap) a é + 6 5 ; 8 ~ a) Da ainzyelsdmay ni G f s e = E FE 5 = i o 5 A : = < 4 . . : . . * : K : 5 4 4 2 ae i re oe 5 5 5 5 ; > > > > > 8 Pee b 3 8 Gj 6 6 ° ° ° Z Z a a 4 @Lake Wingra froze over on Nov. 15, but opened up again; it did not freeze over for the winter until Dec. ro, 346 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 5.—PERCH CAUGHT PER HOUR AND PER Day in Gi Nets, LAKE WINGRA, 1916.4 May 13... Gill nets. Perch caught. Date. Size Average | Average mesh Time set. | Time pulled. per per (meters). | (inches). hour. | 24 hours. BPN Be scccbesscndesasatt % 7-00a.m.| 11.00 a.m. 4 5 13-7 329 I 7.00 a. Mm. II.o0 a, m. 4 3 7. 18 2 7-00 a.m. II. oo a, Mm. 4 ° ° ° I II. 45 a.m. 2.30 p.m. 2.2 3 1s 360 2 Ii. 45 a. mM, 2.30p.™, 2.2 ° ° ° BHOTIITAS Choo ic os Sbie taice Baile G4 ANT jhe. ee ce oe ow calelb ecto ea eey on ns 8 16 384 BOI 20.8. bis Mic daicte ct uM 7.00 a, mm 8.25 a.m. m3 26 17:3 415 14| 6.s0a.m II.oo a.m 4.2 ° ° ° +7 2 6. 50 a. m. II. oo a. m 4-2 ° ° ° I I 6.50a,m 11-00 a, m, 4:2 4 “9 a2 BOE. B20. baa diedeonttech|s 3 1%4| 7-10a,m. 7-20a.m. 2 ° ° ° 5 % 7.10a, m. 8.15 a.m. It 6 15 360 I 4, 7.15 a.m. 8.50 a. m. 1.6 ° ° ° I I 7-20a.m. 12.00 Mm. 4.6 ° ° ° 1-5 4 9-00 a.m. 12.00 1m. 3 ° ° ° 5 “% 8.ooa,m.| 11.30a,m. En ° ° ° rs I 8.00 a.m. 11.30 a, m. 35 3 “9 22 5 I 8.coa.m.| 11.30a,m. 305 ° ° ° Ys) Bee nag IO bOnS Abd Mine aeceee 2 I II. 45 a. Mm. 1.00 p.m. re4 ° ° ° 2 1%4| 7-45 a.m. 8.10 a, m. 6 ° ° ° I % 6.35 a.m. 8.15 a.m. 1.6 9 5-6 134 2 I 6.35 a.m. 8.50 a.m. 2.3 ° ° ° 2 14) 8.10 a.m. 10.15 a.m. 21 ° ° ° 2 % 8.10 a.m. 10. 15 a, m. 2.1 ° ° ° 4 7-45 a.m. 8.00 a, m oe ° ° ° I 7-45 a.m. 8.00 a, m re ° ° ° 144| 10.308. m. II. 45 a. m. I.2 ° ° ° I 10. 30a. Mm. II. 45 a. Mm. 2 ° ° ° 144] 10.30a.m. II. 45 a. m. 2 ° ° ° I 1o.30a.m.| 11.45 a.m. 2 ° ° ° MAY G25. Gs siaiaccljacinethes % 6.30 a, m. 9-35 a.m. i 42 13-7 329 I 6. 30a, m, 9-45 a.m. ° ° ° 4% 6.30 a. m. 10.00 a. Mm. ° ° ° I 10.15 a.m, II.oo a. mM, 4 5.2 125 14] 10.15 a.m. 12.00 m. ° ° ° 14] 11-30a.m. 12.00 m. ° ° ° I 12.00 m. I.00 p.m. ° ° ° I 10. 00 a. Mm. II1.coa.m, ° ° ° 144| 10.00a, m, II-o0 a. m. ° ° ° I I. 00 p.m. 1.35 p.m. ° ° ° %4 1.00 p.m, 1.35 p.m. ° ° ° 4 6.45 a.m, 9.00 a.m. ° ° ° Ir 7-00 a. Mm. Io. 30a. Mm. ° ° ° % 7.00 a.m, 9.00 a, Mm. 3 6.5 156 I I1.30a.m,| 12.45 p.m. ° ° ° 14| 10.30a.m. II. 30a. Mm. ° ° ° 14| 11-3048. m. 1.00 p.m. a ° ° ° I 12.45 a.m. 2.00 p. Mm. 3 9 7-1 170 1%| 12.45 p.m. 2.00 p.m. 3 ° ° ° May aosscastosctecsecinee 1%4| 6.474. m. 8.30 a. m. 7 ° ° ° 14| 8.40a.m. 9.12 a.m 5 ° ° ° 14 9-12 a.m. 9.28 a.m. +3 ° ° ° 14| 9.28 a.m. Io.50a.m 1-3 ° ° ° I 6.53 a.m. 8.35 a.m. 1.6 ° ° ° I 8.43 a.m. Io. 40 a, M. 2 ° ° ° % 6.58 a.m. 758 a.m. I 16 16 384 % 8.45 a.m. 9.18 a.m ae 5 10 240 % 9-35 a.m. Io. 30 a. M I 13 13 312 MAY ia caaicaiednve sien eae ed 1%4| 8.00a.m. 9.30a,m is ° ° ° 1%4| 6.45 a.m, 8.00 a. m. 1.3 ° ° ° 14 9. 30a. m. II.ooa.m. rae ° ° ° I 7.00 a, m. 8.30 a. m. 5 ° ° ° I 8.30 a.m, To. 30a. m. Ks ° ° ° I Io. 00 a, Mm. I1.30a.m I5 2 I.3 3t I Ir. 30a. m. 2.00 p.m. 2.5 ° ° ° % 7-20a.m. 9. 00 a. TM. 6 63 37-8 907 VRS y Srnassoeondacnagad pastcscees % 6.53 a.m. 7-55a.m. 21 ar 504 uy 7-55 a.m. 9. 10 a. m, Ir 8.8 211 % 9-10 a.m. 9. 30a. Mm. 4 12 288 % 9- 30a. m. 9.524. m. 3 9 216 I 7.00 a.m. 8.05 a.m. ° ° ° I 9-15 a.m. 9.45 a.m. ° ° ° I 8.15 a.m, 9-15 a.m. ° ° ° 1% 7.25a,m. 8.10 a.m. ° ° ° 4 8.10 a.m. 9.25 a.m. ° ° ° 14| 10.308. m. 1I.30a.m. ° ° ° % 10.35 a.m. II. 25 a.m. I I.2 29 I Io. 40 a. m. II. 20 a.m. ° ° ° @In all nets the mesh sizes were bar measure, The 1-inch mesh nets were 60 feet long; all others, 75 feet long. HABITS OF YELLOW PERCH. 347 TABLE 5.—PERCH CAUGHT PER Hour AND PER Day in Gu Nets, Lake WINGRA, 1916—Contd. Date. AIRING 20. Heptdicis x.» eeidadecaius RINGER Fc iebea steve. ¢ TRY aaah wah ic caecc guess Say AG race 005 van Gee's BSE Son chen hss 40-Oie << oa pee eT ek i055. soancanacusncete CE xx on caraak CE Kas ceceas ees avenue -+| 110307°—21 Tem- pera- ture (degrees centi- grade). 20.5 30 29-4 26 13-7 Gill nets, Depth Size set mesh Time set. | Time pulled, (meters). (inches). 2 1% 6. 43 a. m. 8.42 a. m. I I 6.57 a. m, 7-58a.m. 5 % 7-05 a. m. 7-48 a.m. In5 % 7-48 a.m. 8.15 a, m. I % 5-50 a. m. 7-03 a. m. I % 7-03 a.m. 9-00 a. m. 5 I 6.10 a.m. 7-20a,m. nS I 7- 20a. mM. 9.30 a. m. eS 1%4| 6.32a.m. 7-322. m. Iy5 4 7-32a.m. 9-45 a. m. 5 I Io.008.M.]} 11.30a.m, 5 It Io, 00 a. m I1.30a.m 5 I I1-30a. m, I. 00 p.m. 5 I I.00 p.m, 1.45 p.m. 5 14 9-45 a.m. Io. 15 a. m. 5 14] 10.15a.m 12. 00 m. 1-5 144| . 10.3048. m. 12.00 m. 1.8 4 6.58 a. m. 9-03 a. ™. 2.4 I 6.50 a.m. 9.00 a. mm. 3-1 % 6.40a. m. 8.45 a.m. 2.4 % 7-12a.m. 9. 03 a. Mm. 3-1 I 6.50 a. m. 9-03 a. m. 3-7 14 6.40 a.m, 8.59 a. m. 3-1 4 9-20 a.m. Ir.co0a. m. 2-4 I 9-20 a. m1. II.co a.m. 2.2 14| 9.20a.m,] 11-coa.m. 2.7 % 6.40 a.m. 9-30a. mM. 2-7 I 6.40 a.m. 9-30 a.m. 2-7 1% 6.40a,™m, 9-30 a.m. 2.9 % 9.00 a. Mm. Io.30a. Mm, 2-9 I 9:04 a. Mm, Io. 30 a.m. 3 4 9-10 a. m. Io. 30a. Mm. 35 %| 10.soa.m. 12.25 p.m. 3-3 I 10-53€@.M./ 12.25 p.m. 3-1 14] 10.55 a.m. 12.25 p.m. 1-5 14 1.31 p-m, 2.30 D.m,. 15 I 1.35 D-m. 2.35 D-m. 2-5 x% I.42 p.m, 2.40 p.m, 2 I 8.45 a.m, 10. 45 a. Mm. 2 I It-o0am.| 12.00m, 3-5 I 12-30p.m, I-15 p.m. 2 I I. 40 p-m. 2-40 p-m, 1-5 y% 9.05 a. m. 1o.50a. m. 1-5 | x1.05 a.m. 12-com, 3-5 %| 12.35 p.m. I.20 p.m. 5 % 1.40p-m. 2.40 p-m. x I 7-35a.m 8.40 a.m 2.8 1%4| 5-10a.m. 8.20 a.m, 2.8 I 5-22a.™m. 8.18 a.m, 2.8 x% 5-35 a.m. 8.12 a.m. I % 6.05 a.m. 11.coa.m 1-5 I 6.10 a.m. 11.03 a. Mm. 2 14] 6.15a.m.] 11.05 a.m. I % 6.15 a.m. I1.10a,m 2 I 6.22 a.m. II-00 a. Ml. 3 1%] 6.28a.m. 10.55 a. Mm. 2 x% 6.15 a.m. Io. 20 a. Mm. 2-5 I 6.20 a.m. 10. 25 a. M1. 3 1% 6. 26a, m. Io. 28 a. m. 3 1% 6.15 a.m. 7-50 a. m. 2 I 6.24 a.m. 7-50a.m Tek x4 6.32a.m. 7-50a.m. I 4 8.10a, m. To. 45 a. mM I-5 I 8.14 a.m. Io.50 a.m. 2 14 8.18. a.m. 10.55 a.m. 1-5 x% 6.30a. m. 10.45 a.m. 2 I 6.45 a.m. Io. 42 a.m. 2-5 1% 6.55 a.m. to. 38 a.m. -7 % 6.23 a.m. 11-08 a. m. 5 I 6.324. m. II.04 a.m. 2.5 1% 6.43 a.m. II.coa.m. I I 2.00 p.m. 3-57Dp-m. I I 4-07 p.m, 5-50p.m,. 2 4 2.10 p.m, 3-55 p.m. 2 1% 4-09 p.m. §.50p-m. I-7 I 1-55 p-™m. 5-09 p.m. 2.5 14 2.00 p.m. 5-06 p.m. 1-7 I 5-09p.™, 7-55 a.m. 25 1% 5-00 D-m. 7-508, mM, 23 Hours in water. Cl SUpyeod HHUwO WoO rT » > +5 Hu HHH HNN DDD HHH DDH eb DH UII WRM WOOD AAAW wo I » x BARA E D ice a a eS nee Sewn NOH kVA WH OY Perch caught. Average | Average Total. per per hour. | 24 hours. ° ° c ° ° ° 6 8 192 4 8 192 It 8.8 2Ir aI 10.5 252 ° ° ° ° ° @ ° °o ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° 2 “7 16 78 39 936 Is 8.2 196 ° ° ° ° ° ° 2 2 29 2 1.2 29 ° ° ° 8 2.9 7° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° ° I “9 a2 ° °o ° 3 3 72 32 2-4 26 ° ° ° ° ° ° I +3 7 25 5 120 ° ° ° ° ° ° 35 7 168 a +4 10 ° ° ° 13 3-3 79 ° ° ° ° ° ° ° ° ° ° ° ° 2 1.6 38 ° ° ° 2 8 19 ° ° ° 7 1-7 31 2 5 12 ° °o ° 8 1-7 3r 2 +4 Io ° ° ° x “s 12 ° ° ° ° ° ° ° ° ° 4 1-3 3I ° ° 2 I +07 ° ° ° ° 348 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 5.—PrERcH CAUGHT PER Hour AND PER DAy1N Grit NETS, LAKE WINGRA, 1916—Contd. Tem- Gill nets. Perch caught. pera- Date. cere . ‘ (degrees | Depth Size 4 Hours Average | Average centi- set mesh Time set. | Time pulled. in Total. per per grade). | (meters).| (inches). water. hour. | 24 hours. Ole ae TSR Gide. SATIS SL ARDS 12-3 17 I 7-55 a.m. 7-5s0a.m. 48 I +02 “5 25 1% 7-50a.m. Io. 50a. Mm, 3 ° ° ° 2.5 1% 7.50a, m. 7-45 a.m. 48 ° ° ° «8 % 7-45 a.m. 10.45 a. Mm. 3 6 2.1 50 Bets 205.5 SA ae ccinatuc 8.5 1.6 I 8.10 a.m, 4.00 p.m. 31-8 3 +I 2 3 1% 8.20 a.m, 3-50 p.m. 31-5 ° ° ° 33 2 8.30 a. m. 3-40 p.m, gi.t ° ° ° OOCE AEBS ale ccRaieisinispiteeist I 2 I 8.25 a.m. 8.30a.m. 48 13 3 7 3 I 8.35 a.m. 8.10 a. m. 47-5 ° ° ° 3 2 8. 43 a. m. 8.15 a.m. 47-5 ° ° ° Oct: exit. Bcsocccbatecs 6.5 2 I 8.30 a. m. 8.15 a.m. 71-7 6 +08 2 3 I 8.10 a.m. 8.03 a. Mm. 71-9 ° ° ° 3 2 8.15 a.m. 8.08 a.m, 71-9 ° ° ° 1-4 % 8.05 a.m, 8.43 a.m. +6 2 3-3 79-2 ING: Gee dele occ cece var 7 2 I 8.15 a.m. 8.12a.m. 96 3 +03 7 3 I 8.03 a. m. 8.05 a. m. 96 ° ° ° 3 2 8.08 a.m. 8.00 a, m. 95-9 ° ° ° I.5 I 8.30 a, m. II.10 a.m. 2-7 2 6 14 2 14] =8.32a.m. 11.08 a.m. 24 ° ° ° 3 2 8.35 a.m. 11.08 a.m. 2.5 ° ° ° I % 8.57 a.m. II. 20a, m. 2-5 6 24 57 NOs) Sonic oemclvvsuspecasclcnmaberoen 1.5 I 1I.10a, mM. 6.05 a. m. 19 4 -2 Cy 2 r4| 11-08 a.m. 6.05 a. m. 19 ° ° ° 3 2 11.08 a, m. 6.05 a. m. 19 ° ° ° NOW. Rodos ace scn ber eets 9.2 3 I 6.35 a.m. 7-40 a. m. 49 7 I 3 3 14! 6.35 a.m. 7-40 a.m. 49 ° ° ° 3 2 6.35 a.m. 7-40 a.m. 49 ° ° ° | COAG BORA RSD RAGE Foco7 8 3 I 8.10 a.m. 8.00 a. m. 48 16 3 | 8 3 1% 8.10 a, m, 8.00 a.m, 48 ° ° ° 3 2 8.10 a. m. 8.00 a.m. 48 ° ° ° DNOVc/ EE nicole cle pee vie/e@ienin oie 6.7 5 I 8.40 a.m. 8.05 a.m. 47-5 10 +2 5 35 1%4| 8.408. m, 8.05 a.m. 47-5 ° ° ° 35 2 8.40 a.m, 8.05 a, m. 47-5 ° ° ° I % 8.40 a, m, Io. 40 a.m. 2 24 as 12 1-5 %| 12.00m,. 8.25 a.m. 20. 4 ° ° © 1-5 x% 8.52a,m 10. 47 a.m. 26 I +03 “7 ns %| 10.50 a. mM. 8.06 a, m. 21-3 ° ° ° @Lake Wingrta froze over on Nov. 1s, but opened up again; it did not freeze over for the winter until Dec. 10. 349 HABITS OF YELLOW PERCH. TABLE 6.—Foop oF 499 ApDuLT PERCH IN LAKE MENDOTA, 1915, SHOWN BY MONTHS. [All figures referring to food indicate percentage by volume; + means a trace.} “SIM ps STRIS “6y0e pasui Ay | “syIMpe BIqyaIO | “sympe prarouory5 | ‘zdnd wiqjaiog | : SB Gsp sant | “syued “wav | -zdnd prmonoig | : ” "WAIL peuneeprua | 6: “WAIE] SIEIS | “sydurAd a}e00pO | “waIe] ABSIPped ‘sydurAu puonmeydy *BAI] BIZZaqoIg “BSB snoy WOME] | *sazeIOIYO | “sqpuRIqHeue’y “spodoljsex | "BAI" BIQIIOD “SAIE] prmouolig) “asta | (siazannyyM) Wyug] eeloay Perch. -xa JaqumnN Month. “syoas Month. Average.... BULLETIN OF THE BUREAU OF FISHERIES. 35° TABLE 7.—Foop oF 188 ADULT PERCH IN LAKE MENDOTA, 1916, SHOWN BY MonTHS. [All figures referring to food indicate percentage by volume; + means a trace.]} *s}[Npe e1a}doajop *s[BISAIO EQD¥RD ‘wdnd viyjai05 “sLiqgp pur IIS ‘sydurdu ptisurydy Perch. ‘zdnd prmouomy> *syuuld “apy “BAIL SIBIS *‘sayeYIOITO “sydurdAu 3}e00pO “satjo9e'T ware] AYSIPPED | vt ‘SA1v] pramouoly) Ox gevnrana] Ss ‘sqpuPIqyyaure’y ‘spodosjsexy *BAIL] BITIVIOD “B1DD0PLO aHama *spodados oo SHE 8 “spoors43sO, Pee ee ec ee te *$339 Ysty a Gesiqne—s ° sroqamy [Tar 2 ; on Bee g(a | HEE | 8 saysyAvig -xa as quinn | ‘synpe paynuepmg E ‘ 4 g : : : 2 z FRA rene cic oeerioe 186 | Oligochztes........... 20 30-0 | Sialis infumata larve, o.s; fine débris, 15; plant remains, so. HABITS OF YELLOW PERCH. 353 TABLE 11.—RATE oF DIGESTION OF VARIOUS Foops AT A NEARLY UNIFORM TEMPERATURE, 16.6° C.4 Time in hours until first feces. Temperature (degrees centigrade). Food eaten. Shortest.| Longest. | Average. 16.50.58 5-0 7-0 6.0 16.3... 4:5 6.8 5-4 16.8. 7-16 23-0 19-2 16.8. 5-5 24.0 8.4 16.5. 3-25 23.0 b 135.8 17-0.. 24-0 24-0 € 24.0 @ xo perch were used; minimum weight, 1.9 g.; maximum weight, 4.67 g.; average weight, 2.85 g. When possible, as with chironomid larve and earthworms, the rate of digestion was computed from the time the fish began to eat; in a few cases (Entomostraca), from the time the food was placed in the dish. The experiments extended from Sept. 16 to Oct. 2, 1916. © Not eaten by 5 individuals. ¢ Not eaten by 3 individuals. TABLE 12.—RATE OF DIGESTION OF VARIOUS Foops AT TEMPERATURES OF ABOUT 3 AND 18° C.¢ Temper- Food eaten. Average ature pe A EE ee eee ee ete Date. (dese digestion centigrade). Kind. Number-} (hours), Jan..24,1917...-... 2.5 3 43-7 1 bcd Seaaioee 18.0 16 22.0 Jan. 17,1917. 2-5 (?) 47-5 JRiarssXOr7s. 0. 18.0 (@) 23.0 Dec, 18, 1916....... 3-5 | Dikerogammarus.. 3 64-0 AD f:\s Ot -Pp <-> > Pe 2.5 | Earthworms I 390 10 yee SaaS 17-0 I 18.4 Jan. 22.1917..... 2.5 ° 3 OY COREE CS 7 SEES 2.5 I 43-3 20.5 I 16.4 4:5 aaa _ 42 experiments were conducted simultaneously, using 2 perch in each case, averaging 28 g. in weight. When possible, as with chironomid larve and earthworms, the rate of digestion was computed from the time the fish began to eat; in a few cases (Corethra larvz), from the time the food was placed in the dish. TABLE 13.—RaTE oF DIGESTION OF VARIOUS Foops aT TEMPERATURES VARYING FROM 23 TO 26° C.4 Food eaten. Yemipery! 22.6 3 ed Se es Aerie ature rate of Date. (degrees Number. Volume (c. ¢.). diges- centi- 2 tion grade), Kind. (hours). Total. | Perhour.| Total. | Per hour. {EOS ARCO, GMS: Ber raeeS Sate ree, 23 | Damselfly nymphs............ 24 1-5 2-5 0.15 31-7 July 20, 1916.. = : 26 | Earthworms. ......+..-++6-005 5 2 2.0 +16 20.0 July 5)1916.......+... 23) MIMD WS cit cue ome cea duatt oe 13 +57 5-2 -23 22.5 @1n this experiment 1 perch weighing 48 g. was used. TABLE 14.—RATE oF DIGESTION oF VaRIOUS Foops AT TEMPERATURES VARYING FROM 2.5 TO 4.5° C.2 Temper- Food eaten. Average Date. ature rate of (degrees : digestion centigrade). Kind, Number. | (hours). Jan, 24,1917 2 5s} osronomiid TArVesp sec dere ose cee ta se etna e 50 46-5 Jan. 15,1917. . 2.5 | Corethra larv@............0.000s eh. (?) 48.0 Dec. 18,1916. . 3-0 | Dikerogammarus aete 3 64.0 Jan. 8, 1917... B.6 | ALCO WOkttS. «6 deine nae cos eile sina ona ° Jan. 22,1917. . 2.5 | Liver and flour. . Aree ° Jan. 1,1917... 2.5 | Minnows..... ey] 6 Dec, 26,1916... 4-5 | Snails.... ° In this experiment a perch, weighing respectively 31 and 24 g., were used, 354 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 15.—RATE OF DIGESTION AT DIFFERENT SEASONS, AS JUDGED BY INTERVAL BETWEEN EATING AND FirsT APPEARANCE OF FECES. [Summary of data presented in Tables 11 to 14.] 7 Fish used. Food eaten. stage [=] Avera LES (deerees Average : digestion Number.| weight Kind. Number. } (hours). grade). (grams), Oct. 2,1916 16.5 10 2.8 6.0 Jan. 24) 191 2.5 2 2-8]. 43-7 Do.. 18.0 a 2.8 22.0 2 an cgagnnuce agape 2-5 2 275 46.5 Jerry rors 3 tee ee 2.5 2 2.8 475 Jans us. tonz7 ee ek 18.0 2 2.8 23.0 OAR rotAtitesick. 2.5 2 27-5 48.0 Jume 19, 1916... .......... 23-0 I 48.0 31-7 DeceeS EGG a iesiee res 3-0 2 2.8 64-0 Day Ts. seed Tt 3-0 2 27-5 64.0 SEptssoXOrG sc cewesccvce 16.3 10 2.8 ea Sept.:20, 1916... ee 16.8 10 2.8 19.2 JAUINS AOKI <2 a/c sw niese 2.5 2 2.8 \ Davs\.... 17-0 2 2.8 July 20,1916. 26.0 I 48.0 Jan. 8,1917.... 2.5 2 27-5 Sept. 25.1916. . 16.8 10 2.8 Sept. 22, 1916... 16.5 10 2.8 Jam. 22,1917... 2-5 2 2.8 Dees: Pap.» 25 2 27-5 Sept. 27,1016... 17-0 10 2.8 Jan. 1,1917.... 2.5 2 2.8 Jan. 5,1917.... 20-5 2 2.8 July 5,1916.. 23-0 I 48.0 Jan. 1, 1917 2-5 2 27-5 Dec. 26, 1916. 4:5 a 2.8 June 30, 1916. 24-0 I 48.0 Dec. 26, 1916. : 4:5 2 27-5 GER AO IOKG cscs ccc cen: 17-0 ro 2.8 TABLE 16.—PERCENTAGE OF VARIOUS Foops EATEN BY 715 ADULT PERCH CAUGHT AT VARIOUS DEPTHS AND AT ALL SEASONS IN LAKE MENDOTA DURING rgI5 AND 1916. é|gela]4 ji Depth at which caught b=] e ] 3 g F 5 3 a R (meters). cee pone a a a/s 5 | 3 r a tir t= Me Fis] pra es a Uefa = PRT Ve Yad WPT Ba a a % 9 9 g EY s 4s so} a | So 3 = 1 = :) = @ > 2 & & g = Rael oa | aS | eee pa ae ee | BO) inte) eee | lesa Rico lO QO BORA. bis ane atm asain 5-3| 412] 33-0] 4.0] ro] rH] 7-0] 9-5] 0-4] 2%-2] O4] 3-2] O07] TO 9-5 . 8.7 oS 4 1-7] 5-0 +I at +3 . 1.6] 25 1.6 8.4 eI ° 8 +4 Oo -t 3-4 3-4 ” oo ~ » w& a ro] a ° t a re} uw H w » a ” 2 ” ” ° n wo TABLE 17.—NUMBER OF TIMES A PARTICULAR Foop FoRMED THE LARGEST ITEM, BY VOLUME, IN ALL PeRCcH CAUGHT AT THE SAME TIME AND DEpTH.2 [Percentages refer to numerical ratios, not to volume of food.] Cla- Oligo- Depth (meters). ish. 1 doc- | Snails. Bvieeres! era. @ 687 adult perch caught during rors and 1916 in Lake Mendota are included. 355 HABITS OF YELLOW PERCH. = ‘qduAu co: 5 WINjeUs}Ue eUse[ [eu : Aa & g *ydurAu tuesey eudeyeng | aE é “BAI abd = eyepundyinur rayAelsy i ° in wz *syduAu eyNuTMp siue_ | = = g a F § “yqdurtu sneg + 2: ° a — Q g sydurdu AgAryy ‘3 m O&A Sie pe et “syduAu Appsueq a e # — 2 ait. eae Bo 3 WAIL] VUIISSI}NUTUL val éiat & : = Pr ty . ewe Oe eS "WAIL] sl[wour sndAuey, seni ae § ze < + ar ce] g § *BAIP] Pas 4 Oy, vg || snjerojooap sndAuey |' ©: : mw ¢ — asi & | Tees a 2 “BAIL] sNIpepPoqUuO Dig ae 8 Beate Mid aoe: Z 54 “BAIR sHIpeporg 6: 328 1@) Ss “BAIL SNZEWSIp “D ne S| moo : ¢ 4 oHa & 5 “BATE SUPWOAATY “D ] A aS a 2 nin 3 Hl “ware “ds snuouoig) | Kold@ = a n = *(sIojour See ee: < a -I[!ar) que] adeaay BNO OR ° ie =). "aes x a Pa *pautuiexe JaquinN ° : * g g A Ae 5 Date. Fea T airs Sarat os olnge July 19 Aug. 7.. * FBP ceadec A ooh betes PG VA selene o}s si biormib nro pte | saa A TABLE 19.—Foop oF SMALL PERCH COLLECTED ALONG NorTH SHORE OF Oconomowoc LAKE, 1916. [All figures referring to food indicate percentage by volume; + means a trace.] “syduAd =e TAnj}eUte}ue PUMe[] eB Oy ‘sydurdu Apjesureq: *sydurdu Aguoseiq *syduiAu eynurmp suze} “BAIL] Joqe[s eizzaqoig *BAIE] snjeroseyri} sndojoorng “WAIL sNIpepIorg | "SAIL] snsivjAuey “BAIP] siimour sndAuey, SNTO BAIL] [oie} umouyun ‘ds sndAuey aa “wale eiAmodjeg "BAI sNpUNOATU "CO “BAIPT SNIPEPOY UO “WAIL SIVUDALATNY “D “BAIL SNOIOFIGOT “Dd | “Bair ‘ds snomouoig) | “ ~”m ‘BAIL 2 a aie [ Pesuy | aie ss | a *(siaqjour CO: btn Go | -Mim) yuaaseiay | SSSss x Py *ponturexe JequinNn | Date. TYSox g5293 Seaaq ae um “wale snoyOOMTE]T Ty | C “esd ‘stmolizyds sniopAq) ‘smuedoby es a : -emosoneyderqy | Baga 8 *snxoma[ qd :3 onn ” ‘eruqdeq | .. .. -- tak) F no: Sn}e][eute] snosoAIN Gy aise *Bynu09 tn ” STI}SOIIZUO] BUTMISOg 4 “erydepora) mouse ov ° sdopsy] gg ick me *spoors}sO 3 ip 4 ™O MOO + “B099Ze ELayeAHL | é3 < 53 ® “saysyAeig les 78 ‘zwdnd erAmodeg | Pit :° -wdnd prmouoimy) | “BAIL vaAvisy “wasey AYSIpped WAIL] VIVANT sIPLIS Date. 356 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 20.—Loss AND GAIN IN’ PERCH FED DIFFERENT Foops From AUG. 19 TO SEPT. 18, 1916. Aug. 19 Sept. 18. Per cent gain. Per cent loss. Number Average Average Food. of perch Volume Volume feat’ ‘Average | volume | Average | volume Weight || (cabic | Weight |) (cubic roe (cubic eons (cubic (grams). || centi- {-(grams). | centi- grams).| centi- grams).| centi- : ‘ meters). meters). meters). meters). Liver and flour. & J... i5..)..5-..). a3 2-104 2.40 2.50% 2. Hyalella.......... Ns 3 1.856 1.76 2.2 2. Entomostraca. . 3 2-683 2-91 3-874 3 Earthworms.... 3 1.856 1.68 2-726 2. Insects, adult... 3 2-152 2-07 1.643 Ty Chironomid larve 3 2-099 2.00 2.836 2. VCE Hoag eee 8 3 1-944 I-91 2-380 2. Normal. . 2 3-232 3:20 3-850 tS SUSCV esi discnisndevinan chasderae b3 1-205 1-26 -86r @1 died Sept. 6 on account of fouling of water. bx died Sept. x1. TABLE 21.—COMPARISON OF Foops EATEN BY ADULT PERCH IN LAKE WINGRA AND IN LAKE MENDOTA, SHOWN By MonrTHs. [The figures are percentages by volume and + means a trace; 350 perch were examined in Lake Wingra from March, 1016, to February, 1917. In Mendota, 499 were examined in rors and 188 in r916. To obtain averages given for Lake Mendota the figures for r9r5 were multiplied by 3 and averaged with those for 1916.) LAKE WINGRA. a . . g . P 2 3 oe & 8 a g 5 B/E] 8] § Fi el leah fal a t|2i3/% Month. | 2 | = nile bs 3 g a 3 3 5 Bee | hes ce aliaas e a B11 8392 a | @ 4 S 3 s 8 | 2 a6 aye (ieee ee ee et ae ye ie | gy ot = a a Oo = AOL: i WR AO PLP OO a 8 January..-| 41.5 3:0 February -| 35-3 i March.....] 8.7 April, = ia iL Maw 3b s) : 42 38. 26 61.31 (De | Geen beh ; 65 13-74 85.62 13 2minutes.... I-at 7-87 90-92 Xo ORs tsicn 5 minutes..... +4 11.06 88.55 X35) Bl steteles 8 minutes..... +43 14-69 84-88 5 heed eae do. yo minutes... . aI4 12-3 87-36 13 1.5 hours 211 24-24 75-50 13 d 31 19-66 80. 03 13 gr 22-51 77-18 13 Ir 17-90 81.99 AT 2 yo ae SAE Ae (>) 64 18. 51 80. 85 13-5 i 12 +73 99-15 13-5 ..do. 9 minutes. 170 4.08 95-74 Aatgsr. . obs seen and (Dig Tae Se Se i Re oh me 21 8.0 or. 78 13-5 | 1.1 hours 1 minute +43 37-06 62.50 ELH eee do 8 minutes +24 5-8 93-96 13-5 | 2 hours 4 minutes 21 417 95-62 T3eiS tl doses b. .| 6 minutes 044 28.21 71-75 TZ A15 || - 1 se Ge -| 9 minutes. os 27-09 72-86 13-5 12 minutes. 26 20. 51 79+ 23 Hepteasisees ae seaay OP tee | Oe Obmrartiondece) Gana weer 33 28.24 91-43 2 89 25-93 73-18 2 62 34-12 65.26 2 71 25-65 73-64 2 24 32-43 67-33 2 25 25-06 74-68 Septerahe co sodee act 3 2673 28.65 68. 63 Vc all Reae pa Sele HP 78 29- 61 69. 61 13-5 1.20 24-78 74-02 433.5 +25 25-31 74-44 13.5 +43 9-96 89. 61 433.5 3-0 14.19 82.81 433.5 |.....do.........| 8 minutes..... 1.09 II-20 87-71 @ Some individuals had been near the surface of Lake Mendota; others had been in the stagnant water below the thermocline, where there was practically no oxygen. > Surface. ¢ These individuals were caught at a depth of less than 2 m. in University Bay. d@ These individuals were kept on the deck of a boat in stagnant water pumped from 13.5 m. — HABITS OF YELLOW PERCH. 359 TABLE 24.—PERCENTAGE OF GASES IN THE Swim BLADDERS oF PERCH.? Respira- Num- | Instag- tion Depth ber nant |Average| Most Least |Average| Most Least |Average| Most Least quo- (meters). of water CO: CO. CO... On. Or O2. Nz. Na. Nz. tient perch. | (hours). CO. ) @ Some individuals were examined as soon as taken from surface water; others, after being left in the stagnant water below the thermocline. ‘TABLE 25.—S1ZE OF PERCH AT SEXUAL Maturity In LAKES WINGRA AND MENDOTA, MARCH TO OCTOBER, 1916. Perch examined. Size of perch (millimeters). Lake. Number. Sex. Maximum.| Minimum.| Average. 162 | Male...... 178 113 134-9 158 | Female.... 180 118 137-7 74 | Male...... 187 115 156-6 95 | Female.... 201 120 167-6 360 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 26.—SEX OF PERCH EXAMINED FROM Various DEprus or Lake MENpDoTA Durinc 1016. Date examined. Depth (meters). HH ne PyG Spoga eisiu & ooo~soounodouunn > ws OUNONOONN ©COCOOOO0COFROHOOWDDD0O00000000HH Males. w » AUKWOKWOHKYTO MN OOO 4 u H CO0OHFOnRNO Hh OOmMnHOhBKHOOCOOCODCOOO0COO0000CO00 @ These nets remained in the water 14 hours; all others, 3 hours. ©COCODD0OHHOHHOOCOCOODOOHODOOCO OS Females. COCOOKDDDDODDODOOOO OH OMT ONWOWH ww OnASS OOS a HY DOHUVU HUE HUNOO TABLE 27.—FisHES CAUGHT PER Hour ON HOOKs AND LINES BAITED WITH MINNOwS, LAKE WINGRA, 1916-17.4 Temper- ature Number 5 Perca | Pomoxis ;. | Leposos- | Microp- | Eupo- Date caught. (degrees | ( bred _ of hooks te nite) fal- spa- | LePomis|““teus |terussal-| smuotis centi- | MEETS.) “used. ours). | vescens. | roides MCISOF. | osseus. | moides. | gibbosus. grade). 20. 5 3 2.2 ° 23-6 O4 Oo. 4 ° ° 30-0 3 32 ° 207, ° ° ° ° Bice Daca. 4 ma 25-0 33 20.8 ° ° ° Al ue era catered 4 33 4:5 +9 12.1 ° ° ° 29.4 2 1.4 2.8 ° ° ° ° ° Aan) Fear 2 2.5 ° 1.2 8 ° ° ° state fatten oprasec' Cy 4.0 5 ° ° ° ° ° 27.0 I 2:4 ° 4 ° ° ° ° 26.8 3 3 ° ° ° ° ° ° dad bonAcason loo 3 2.6 ° +3 ° ° ° ° 19.9 |. 3 9 ° 34 ° ° ° ° 14-5 |. 3 34 32 ° ° ° ° ° 13-5 |. 3 27 2.2 ° ° ° ° ° 14-8 |. 3 37 ° ° ° ° ° ° 14-2 |. 3 2-9 1.0 ° ° ° ° ° 13-7 |. 3 -6 1.6 ° ° ° ° 1.6 12.3 |. 2 2 2.1 ° ° ° ° ° 6.5 |. 3 +7 4 ° ° ° ° ° 7-0 |}. 3 1.8 2-7 ° ° ° ° ° 6.7 3 1.0 ° ° ° ° ° ° S56 3 ag ° ° ° ° ° ° 3 I.2 ° ° ° ° ° ° 3 2.7 ° ° ° ° ° ° 3 3 ° ° ° ° ° ° 3 3 ° ° ° ° ° ° 3 +6 ° ° ° ° ° ° 2 I 2 ° ° ° ° ° 2 I ° ° ° ° ° ° 3 2.5 ° ° ° ° ° ° 5 1.3 ° ° ° ° ° ° 2 2.1 25 ° ° ° ° ° @ The lake froze over on Nov. 15; opened on Nov. 28; froze over again Dec. ro. HABITS OF YELLOW PERCH. 361 TABLE 28.—PERCH CAUGHT IN THREE 1-INCH MESH, 3 BY 60 FEET, Guu, Nets, LAKE MENDOTA, AUG. I2 AND 13, 1916.4 ee ses ——— 2.9m. 7-5 m.¢ Surface.@ Time net in water. Number. Time net in water. Number. Time net in water. Number. 8.35 a. mM. tO 12.33 PD. M........20-5 €1}9.12a,m,to118p.m € 112] 9.00a. m. to1.0cop.m....... ° 12.33 p.m.to 4.30p.m, “- 3|118p.m.to5.18p.m... 82 | 1.00p.m. to 5.0cop.m... ‘ ° 4.30P.m.to 9.cop.m, 1} 5.18p.m. tog.18 p.m... 71 | 5.cop.m.to9.10p.m....... ° 9.00p.m.to 1.208. m. 1] 9.18p.m, to1.43 a.m... 24 | 9.10p.m, to1.32a. mM... ° 1.20a.m.to 5.00a. mM. 1 | 1.43a.m.tos5.18a.m... 72 | 1.32a.m.to5.10a.m. 12 5.00a,m,to §830a. m. 1 | 5.18a. m. to9.18 a. m 76 | 5.10a. m. to9.00a. m. ° @ Thermocline at 8 to 9 m.; strong northeast wind. b Set on bottom near shore. ‘ ¢Set on bottom out in the lake. d@Set above the net set at 7.5 m. ¢ Taken ashore; all others were thrown back as soon as they were removed from the net. TABLE 29.—PERCH CAUGHT IN THREE 1-INCH MESH, 3 BY 60 FEET, Gi. Nets, LAKE MENDOTA, SEPT. 7 AND 8, 1916.4 2.8m. 9.5 m.¢ Surface.d Time net in water. Number. Time net in water. Number. Time net in water. Number. I} 9.498 m.to 1.52p. 18 1.52p.m.to 5.48p. 39 5.48 p. m. to 10.08 p. 12] 10.08p.m.to 1.44a. 1.47a. m.to5.38a.m... 13 1.448. Im1.to 5.48 a. 5.38 a. Mm. to9.43 a.M....... 19 | 5.48a.m.to 9.55 a. - 9-43 a. mM. tO 1.43 Pp. M....... Oe haeeenetnvardia: che uisia'pwiMualayy. aininiaeye/dd |e aetute fnvalere 1o.10a, m.to 2.04 p.m. 2.04p.m,.to 5.55 p.m. 5.55 p.m. to 10.18 p.m, 10.18 p.m.to 2.13 a. Mm, 2.13a.m.to 6.03 a. m. 6.03 a. m, to ro.10 a, m. 9. 1.43 p.m, to 5.37 p.m. 5.37Pp.m. to9.47p.m... 9.47P.m. to1.47a. mM... 43 a. m, to 1.43 p.m.. 7 BREESE O00 ONnw @ Thermocline at 10.5 m.; strong northwest wind decreasing to moderate. > Set on bottom near shore. © Set on bottom out in the lake. @ Set above net set at 9.5 m. TABLE 30.—PERCH CAUGHT IN THREE 1-INCH MESH, 3 BY 60 FEET, Git, NETS, LAKE MENDOTA, SEPT. 9 AND I0, 1916.2 2.7m. 11.2 m.¢ Surface.¢ Time net in water. Number, Time net in water. Number. Time net in water. Number. 11.45 a. m.to 4.00p.m. o/} 11.158. m.to 3.07p.m,.... €60 | 11.258, m.to 3.25 p.m.....! ° 4.00p.m.to 7.42 p.m. o}| 3.07p.m,to 7.15p.m..... 67 | 3.25p.m.to 7.33p.m..... ° 7-42 p.m. to 11.32 p.m... ° 7-15 P.m, toz1.15p.m..... ° 7-33 P.™. to11.20p.m..... ° 11.32 Pp.m.to 3.27a.M... o| 11.15Pp.m,to 3.06a.m,.... o| 11.20p.m,to 3.15a.m..... ° 3.27a.m.to 7.37a.m... o| 3.06a,.m.to 7.c6a,m,,... €39] 3.15a.m.to 7.20a.m..... €rx 7-37 &. M. to 11.45 8. M o| 7.06a.m,.torrmi15a.m.,.... 29 | 7.208, m, to11.23@.m..... ° @ Thermocline at 10 to 11 m.; brisk south-southeast wind. b Set on bottom near shore. ¢ Set on bottom out in the lake. d Set above net set at 11.2 m. ¢ Taken ashore; all others were thrown back as soon as they were removed from the net. 362 BULLETIN OF THE BUREAU OF FISHERIES. TABLE 31.—PERCH INFECTED BY CERTAIN PARASITES IN LAKE MENDOTA, 1915-16, SHOWN By MONTHS. [Numbers in parentheses indicate percentage infected of total catch for a single month.] Perch. Parasites. Tape- Adult Larval Month. Num- Num- Nema- | Trema- | worn tape- proteo- ber ber _todes | todes | cystsin | worms | cephalids| Acantho-| Leeches ex- in- in intes- | in intes- | liver, per- in in cephala. | (Piscola). amined. | fected. tine. tine. |itoneum,| intes- intes- etc. tine, tine. 91s. March: .csrcossresicrsieereerest 4t ane} aweeete tire 11(27) 20(50) 11(27) 4(10) fas} asslaiare woke a TE cacimsnimnaeaiiicn tic eeewes 28 24( 86, abo Pe AN 15(53) Maye actsacnanusGgccuneeeweiae 45 30(67) 10(22) 1(2) 16(35) MIMEL cawtcicnascaoite saw ctseerane 60 28 42 LACAN S crime cote 16(32) oe tiantionsc he sickwiew Sencar alee 66 44( 66. 13 a es 42(96) AT gtist:, niece oem coer ote ere Ook 57. 43 2s) 10(22 11(25) 29( 72 PSS PPETT DOL sn alate mirie iets tie nine io nlaivis 43 38(90 35(87 October ssc cercdekiepecsscewanere 53 48(90) ; Fer -| 47(92 November i... s1.5f-4..5. «288Gb «lek 37 36(97) ARONA. ocak s.- 32(86, Decent ber sc cinarcisiosiccktnest emene 36 2n03s) los. ob aa 1(3) 20(55 STORE Spastic tare foiniasaaiactiie 46-6 | 34-5(76) | 7-1(14) | 2-5(3-8) |27.2(57.8) | 2-5(s-2) | 1-5(4.2) | 4-0(9.4) o.1(.3) 1916, Ue fl See daadaeeineaoucctgodcd 10 16 C-)) BSE 1(10) MeBraATY Sorc toe eas nicole ae 10 10( 100, 1(10) 4( 40, a asa 8. is, ape tate wwitayd al axcheye Mbayol svete acts 533 BUD LOE Pay ate alc tacnps or stage create isto wisiesausrets « evenalaic viata. ciara: oot avacelase svacs\s, hey guava. danislapecatehans’ syagoretarecstaasters 539 RlOssatyee state ce ere vere renee ic crctnnetey netcha ts eimneeistth Whe “Reis a aiveiehaneeeejeraie al slulerevacalotte ce eet mete aes 548 (GK PIAMALIOM\OL PIALEA sac sees eictareoelcs oreitlo mse ave cimitesatthe aioe elevate te arya cisterceatet tna are TMCV berate: came ea 554 Sree tpt sista siore 5 Syvatopesc ar etate ors oral ole eter te oWterans onedpioler ass eivinve's clave: tovevarelstateyetstate aieretann «Lins nints Ripe ete wee I FORM .2THSTAOD a Tew | ty nen Os “ pation platen + > Mame! cues Skin Sate bates a4 e144 yondiey n tof thd Bis TeHiaR OLF A: Rt Wad) ssi hes nes “4 Mawr feu sage franeet he Read ome ‘ t a hl ) Bs So hl : eae EAM sity Yo Irene ie: : se it ‘roluregit So Wal fy hse tien pile tt i at a oe Ada vata Sleietnl Tenoo ey 4 vee se Bhat Dray to dpaty | vgs a rat oats ky vw Fwitonat te “elt Sed bee nee ee cael seta. emi maT We doen) Ae ae, iatle i, bry rained at misses Reso IO REE cori st oy ag ey Ge rate, 2 aan mse che eth sree eae bEe ‘ } ahaa | stoolilleolt ie oy in to a« ‘aidbeatl Smeyemynny to MARGIT ne t . lantoansd a lola: nee : “Tether : Ls by 4 : MELD Peete “dee \ iaelas ‘acliekchinns baw va Aaiher. ge eh iM ye oo nays ea es, WO eat Dite oi aa ge 2G. Ip tntunecr 8 oitamtaby es lt na! rh +> SEs NO ita footiiasbl ‘ebiodl: fe anirgnsents bar “mei naa” Tis Masel OW mire Meras rf el is MARINE ALGA OF BEAUFORT, N. C., AND ADJACENT REGIONS. z By W. D. HOYT, Professor of Biology, Washington and Lee University, Lexington, Va. ad Contribution from the United States Fisheries Biological Station, Beaufort, N. C. * INTRODUCTION. It has generally been believed that the greater part of our Atlantic coast is barren in respect to an algal flora. Although a number of species have been recorded from Norfolk, Va., and Charleston, S. C., and a few isolated collections have been made at points in North Carolina and elsewhere, Johnson (1900) recording between 25 and 30 species for Beaufort, it has generally been held that from Long Island Sound to Florida few individuals or species of alge are to be found. The reason usually assigned for this sterility has been the supposed lack of places suitable for attachment afforded by the sandy coast of this region. While this belief is justified for the greater part of the area, the present studies have shown that it is not warranted for the entire region. One hundred and forty-two species and varieties have been observed here by the author, all but 10 of these being found at Beaufort. While this number is not large compared with 525 recorded for New England and 744 reported for Great Britain, a single locality yielding 132 species and varieties can not be called barren. The area included in these studies extends from Ocracoke, N. C., to Tybee, Ga. (lat. from 35 to 32° N., map 1), but by far the greatest part of the work was done at Beaufort, N. C. (lat. 34° 43’ N.), only occasional visits being made to other localities. Studies in the region of Beaufort were made at the United States Fisheries Biological Station at that place from June or July to September or October during the years 1903-1909. ‘Trips of a few days’ duration were made to Beaufort by the author in May, 1907, and April, 1908, and monthly collections of all species observed were made by the laboratory staff from November, 1908, to June, 1909. Visits to regions other than Beaufort were made, as follows: Ocracoke, N. C., August, 1907; Wrightsville Beach, N. C., July, August, and September, 1909; Southport, N. C., Georgetown, S. C., and Pawley’s Island (near Georgetown), August, 1909; Charleston, S. C., July and August, 1909; Port Royal, S. C., and Tybee, Ga., August, 1909. In addition to these alge, the author has studied two small but interesting collections made by Lewis Radcliffe on the coral reefs offshore from Beaufort in August, 1914, and several collec- tions made offshore in this region, principally in the vicinity of the coral reef, by the Fish Hawk, in July and August, rors. 37% BULLETIN OF THE BUREAU OF FISHERIES. 372 (-}1U] O1oqnoseyy jo y3J0U sayrur Sx ynoqe ‘Ja[uy sasooy_ Jo qINOs YsNf ‘fauUEYD syUL_ Ul apeur aJaM OUI sty} UI suONDaTIOD ay, “dear aq} WO pasiaAas 34 P[NOYs Towa a[IAs}qsUM\ PUe Ja[U] O1OqQuOsE]Y Jo suONFOO] ay.) “APE aJaM SUOTPaTTOO FOIA 3 SoIgT BIO] SULMOYs ‘sarpnys aseq} Ut papnpout uoay—' | av S WH F901 MARINE ALG OF BEAUFORT, N. C. 373 In the preparation of the present report three objects have been kept in view: (1) Only occasional collections have previously been made on the coast of our southern States. While the algal flora of New England and Long Island has been studied with some thoroughness and the Florida coast has received considerable attention, the area between these regions has been almost untouched in recent years. Although the present work makes no pretense of being a taxonomic contribution, the effort has been made to present as complete an account as possible of the algal flora of the region, with remarks on species affording interesting comparisons with the same species found in other regions. (2) Little is known of the conditions of algal growth and of the factors limiting their distribution in space and in time. Notes have been made on the con- ditions observed at Beaufort, and some interesting effects of these conditions have been recorded. It would be desirable to have a detailed and thorough study of the con- ditions made here. (3) No work suitable for American collectors who are not trained students of algee has appeared in recent years, and no such work has ever been written for the algze of our southern coast. Although this lack has been partly filled by Collins’s ex- cellent treatment of the Chlorophycez (1909, 1912, 1918) and key (1918a) and Miss Til- den’s work on the Myxophycez of North America (1910), the need still exists for a special account of the alge of this region. With this object in mind, the present report has been written as simply as possible. Technical terms have been avoided whenever the meaning could be expressed otherwise without too great circumlocution and without sacrifice of accuracy. Nearly every species has been illustrated by a photograph or drawing, since an illustration will often give, to one not a special student of the alge and even to the trained algologist, a better idea of the species than pages of description: Two keys have been prepared, one (an artificial key to genera) based as far as possible on superficial, easily observed characters, the other (a natural key to divisions, orders, etc.) showing the diagnostic characters which warrant the placing of the different forms in their respective groups. All photographs and, except where otherwise stated, all drawings are original, nearly all the photographs being made from living plants and all the drawings being made with a camera lucida. In the descriptions of the various groups and in the natural key free use has been made of current works, especially those of Engler and Prantl (1897-1911), De Toni (1889-1907), and Collins (1909, 1912). The descriptions of the species, however, are based in part on specimens observed by the author, includ- ing those found at Beaufort and those in American herbaria which were accessible to him. In using the artificial key to genera and the keys to species, it should be borne in mind that these have been prepared for the particular genera and species mentioned in this work, and if used for alge of other regions may lead the student astray. Even in this region these keys may cause confusion if genera and species other than those mentioned should be found. A collector should, therefore, always carefully compare his specimens with the descriptions before venturing to assign them names. The gross measurements of the size of species should not be taken too strictly, the figures given being the limits of specimens observed by the author or for which a record has been seen. It will be noticed that the descriptions of many of the species are incomplete in that no mention is made of male plants or organs. This is due to our imperfect knowledge of these plants, since, partly because of their inconspicuousness and partly because of their greater scarcity, male plants and organs have been studied much less 374 BULLETIN OF THE BUREAU OF FISHERIES. than have the other forms of plants and organs of reproduction. Svedelius (1908, 1912) has shown that, in Martensia and Delesseria sanguinea, the male plants have an exceed- ingly short duration, in the latter species not more than one month. Miss Dunn (1917) has called attention to the fact that, in Dumontia filiformis on the coast of Maine, the male plants are found only during a few weeks in the spring. A similar scarcity of male plants has been observed by the author for many species at Beaufort. In spite of extensive searches for them, no male plant of Gracilaria confervoides has been observed, and none of Gracilaria multipartita has been found in the harbor; only one male plant of Hypnea has been found among the hundreds examined; and male plants of Chondria are rare. Many other instances of the same kind might be given. While further search might show these to be more abundant than is indicated here, it seems to be true that, with the exception of a few species, male plants and organs are much scarcer than are the other forms of plants and organs. Because of this fact, anyone finding male plants or organs of a species in which they are not described in this work, should save these for study, or should send them to some other student of the alge. Among the Phaeophycee and Rhodophycee all determinations have, as far as possible, been verified by comparison with type or authentic material. Among the Myxophycee the determinations have been made entirely and among the Chlorophycee they have been made largely by Mr. Frank S. Collins. Under each species references are given to the original place of publication; to the most recent general account of the alge, the Sylloge Algarum of De Toni (1889-1907); and to the works of Harvey (1852- 1858), Farlow (1882), Collins (1909, 1912, 1918), and Miss Tilden (1910), these being the publications of a more or less general nature dealing with North American alge. Ina few cases other references of special interest are given. Citations are given, also, to the two principal sets of American alge, the Algae Americane Boreales Exsiccate (A. A. B. Ex.) of Farlow, Anderson, and Eaton, and the Phycotheca Boreali-Americana (P. B.-A.) of Collins, Holden, and Setchell. With some exceptions, where the works cited were not available, all references have been verified. The arrangement used follows, in most respects, that of Engler and Prantl (1897—1911),? except in the Chlorophyceze, where Collins (1909, 1912, 1918) has been followed. The system of nomenclature follows the Vienna and Brussels rules except in the naming of the divisions, where Chlorophycee, etc., have been used. The retention of these names seems justified by usage, conven- ience, and uniformity, and, although not yet acted upon by any congress, seems to come under the principles of nomina conservanda. Those wishing to know more than is given here about the structure of the alge mentioned should consult Oltmanns (1904-5) and Engler and Prantl (1897-1911), where are summed up the main facts about the structure of alge known at the time of their publication. A work of the present nature necessarily has a limited usefulness and should be replaced by an account of the alge of our entire coast. If the present report contributes toward the preparation of the larger work and serves in the meantime to forward the study of the alge of our Atlantic coast, it will have served its purpose. @ While this arrangement is inconsistent and apparently wrong in many respects, we have not yet sufficient knowledge to war- fant a complete revision, and must, accordingly, use it until we obtain more information about the life histories and structures of the various groups of alge. MARINE ALGA OF BEAUFORT, N. C. 375 The author takes pleasure in acknowledging his indebtedness to those who have helped him in the present study. To Frank S. Collins, North Eastham, Mass., and to Dr. Marshall A. Howe, the New York Botanical Garden, he is especially indebted for assistance in the determination of species given throughout the progress of this work and for much helpful advice and information about the distribution of species and about doubtful points. He is indebted to Dr. N. L. Britton for facilities for studying the alge in the New York Botanical Garden and for the use of Plates CX V-CXIX, and to other members of the staff of this institution for assistance during his work in the library there. To Prof. W. G. Farlow, Harvard University, he is indebted for assistance in the determi- nation of species and for the privilege of studying the alge in his herbarium; to Prof. D. S. Johnson, the Johns Hopkins University, for facilities of laboratory and library furnished for the study of the Beaufort alge; to Mrs. Margaret H. Y. Hoyt, for assistance with the drawings used in this work and with the preparation of the manuscript. To all of these and to others who have helped him in various ways the author wishes to express his grateful appreciation of their assistance. PART I. GENERAL ACCOUNT OF THE REGION. LOCATION AND DESCRIPTION OF BEAUFORT HARBOR. The town of Beaufort lies at latitude 34° 43’ N., longitude 76° 40’ W., about 19 km. (12 miles) northwest from Cape Lookout and 120 km. (75 miles) southwest of Cape Hatteras. (See map 2.) South and west of the town stretches the harbor, a large body of water communicating with the ocean by a wide inlet between Shackleford Banks and Bogue Banks. From the harbor near this inlet extend Bogue Sound to the west and Back Sound to the east, separating the mainland from Bogue Banks and Shackleford Banks, respectively. Extending northwest from Beaufort Harbor lies the body of water known as Newport River, with several creeks, receiving frequent inflows of fresh water. A somewhat similar body of water extends northward from Back Sound. The bottom throughout this region is composed of sand, mud, or shells, and offers no conditions favorable for the growth of alge. The beaches of Bogue and Shackleford Banks are flat, sandy stretches. Shackle- ford Beach and the greater part of Bogue Beach are destitute of alge. Alge are, how- ever, frequently found on Bogue Beach for a distance of about 1.6 km. (1 mile) west from the inlet. Here, after storms, are found great masses of algae washed on the beach or lying in the water along the shore. Many of the plants found here, in all likelihood, have been carried out from the harbor by the receding tide; others have almost certainly been washed in from the coral reefs lying offshore, since several species found elsewhere only on the beach were dredged from these coral reefs; while a few species, represented only by specimens from Bogue Beach, may have come from points farther south, some of these being unknown elsewhere north of Florida or the West Indies, and possibly being brought here by the Gulf Stream from that region or from some of the intermediate submerged coral reefs.% @ While species found only on the beach can not strictly be included in the flora of Beaufort, they are treated as a part of the alge of thisregion. This has seemed proper, since it is very probable that some of these have come from the coral reefs offshore, and it is impossible to distinguish between the species that come from these reefs and those that are brought from other regions. Moreover, in view of the algee found on these reefs, it is unsafe to assume that any species observed in this region has come froma more distant point. Such species may be found at any time by collectors here or at other places, and it is entirely possible that some of these, even if they do not now occur at Beaufort, may establish themselves here, either in the harbor or on the coral reefs offshore. These species are included in the total number given for the region, but are enumerated in a separate list. BULLETIN OF THE BUREAU OF FISHERIES. 376 VNITOYVI HLYON YOdUVH LYOANVAd MARINE ALG OF BEAUFORT, N. C. Bui Hourly observations of the current were made by the U. S. Coast and Geodetic Survey on Cape Lookout Shoals Light Vessel from June 7 to September 1, 1912. These showed that, at this place, the mean current, freed from tidal influence, flowed S. 87° E. with a velocity of 723 m. per hour (0.39 knot) from June 7 to July 5, and N. 85° E. with a velocity of 1.372 krm. per hour (0.74 knot) from July 6 to September 1. From this it appears that the Gulf Stream, following the general direction of the coast, has its western edge, on an average, during the summer season, somewhat westward from Cape Look- out Shoals Light Vessel (see map 3), and about 55 km. (30 nautical miles) offshore from Beaufort Inlet. No observations have been made for this region at other times of the year, but the exact location of the Gulf Stream will, of course, vary consider- ably at different seasons and even on different days of the same season, depending on the direction and strength of the wind. Lying offshore are a number of submerged coral reefs (see map 3) which offer some of the most interesting conditions found in this region. These have been described by Radcliffe (1914). The outer reefs lie from about 29 to 39 km. (16 to 21 nautical miles) offshore at a depth of 24 to 28.8 m. (13.25 to 16 fathoms), while the two inner ones lie, respectively, about 3.3 and 6.5 km. (1.8 and 3.5 nautical miles) offshore at a depth of 8 to 13.5 m. (4.5 to 7.5 fathoms). The largest of these, the “ Fishing Grounds,” was visited by the author on board the Fish Hawk in May, 1907, two days being spent there and 22 hauls being made with the dredge over the entire observed reef. This lies about 39 km. (21 nautical miles) offshore, about 22 km. (12 nautical miles) inshore from the average summer location of the western edge of the Gulf Stream, at a depth of 24 to 25.5 m. (13.25 to 14 fathoms). At the time of this visit the observed length was about 1.85 km. (1 nautical mile) and the observed width was about 900 m. (0.5 nautical mile). Observations made by Radcliffe in the summer of 1914 indicate, however, that this reef is many times larger than was previously known. It is now believed to include Station No. 1 (see map 3) extending many kilometers in the direction of New River Inlet and being several kilometers wide. The lower part is composed of old, dead coral masses, hard and densely packed, with the surface fairly smooth, forming a sort of coral rock, penetrated and honeycombed by numerous worms and molluscs. On and in this substratum were found many hydroids, corals, sponges, Gorgonias, Echinoderms, Lamellibranch molluscs, Crustacea, worms and Ascidians, together with numerous alge. Over the reef swam abundant fish, mainly sea bass (Centropristes striatus), the sailors catching these as fast as they could pull them in. The bottom around the reef was composed of sand and broken coral, and, except for one large, apparently unattached mass of Zonaria flava, all signs of life (including the fish) ceased as soon as its border was passed. Although living coral was abundant on top of the reef, there was no evidence that this is growing toward the surface, the depths recorded on the four visits made by the Fish Hawk to this place in 1902, 1907, 1913, and 1914 being almost identical. Some observations made by Radcliffe on board the Fish Hawk in the summers of 1913 and 1914 disclose interesting conditions in the vicinity of this reef. Other reefs seem to be present at various points along the shore, and coral and algw were found abundantly. Over considerable areas at many points offshore the bottom seemed smooth and hard—apparently consisting of rock as smooth as a floor—and bore scattered specimens of alge. Offshore from New River Inlet there was found an 378 BULLETIN OF THE BUREAU OF FISHERIES. abundance of Dictyopteris polypodioides growing in scattered patches separated by sand. This growth was observed partially covering the bottom over an area extending at least 22 km. (12 nautical miles) alongshore eastward from the inlet, and from near the shore to at least 13 km. (7 nautical miles) offshore, at a depth of 5.8 to 11.6 m. (3 to 6 fathoms). The actual area occupied by this growth was certainly larger than this, since the inner limit was nearer the shore than the vessel could approach, and the outer limit was in water too deep for observation. Moreover, it was found in abundance BOGUE INLET . ; CAPE LOOKOUT NEW RIVER INLET \teokour cor BREAKERS NEW RIVER GROUNDS ~~ 4 LOOKOUT GROUNDS NEW TOPSAIL INLET NOL Fr) TOPSAIL INLET rE FASHING-CRI 7 RICH INLET NO.2 a a Ti SOT ESSES ‘BARREN INLET O MASONBORO INLET NOS ' NO 4 Th Vaal Veal cy sata LL y Zo, “pd “. ~~ - =e — — z ed ~ =“ ~~ “4 ~ 3y; » NAUTICAL MILES ucnT ay Map. 3.—Location of known “fishing grounds,’’ mostly submerged coral reefs, offshore from the region of Beaufort, N. C. (From Radcliffe, 1914.) The largest of these, the “ Fishing Grounds,” is larger than shown here, extending from New River Inlet and probably including Station No.1. Algz referred to as coming from coral reef offshore were gathered from this reef. in July, 1915, offshore from Browns Inlet, about 25 km. (16 miles) northeast of New River Inlet (toward Beaufort), and it may extend westward also from New River Inlet. Its presence here is in striking contrast to the barren bottom observed at other inlets and along most of the shore and raises an interesting question as to the sub- stratum to which it is attached. This must be something other than sand, but its nature was not determined. That rock of some sort is present over considerable portions of the bottom is indicated, however, by the observations of Radcliffe mentioned above and by the further fact that there was evidence of corals and alge on the bottom in the Gulf Stream about 70 km. (38 nautical miles) offshore at a depth of about 115 m. (60 fathoms). MARINE ALGA OF BEAUFORT, N. C. 379 Similar ‘‘fishing grounds”’ occur off other portions of our coast. To the south of this region there are listed by Goode and associates (1887, pp. 53-55, chart 15) 13 fishing grounds off the coast of South Carolina, 3 off the coast of Georgia, and 1 off the northern coast of Florida. These lie at various distances from the shore at depths of 13.5 to 35.7 m. (7 to 18 fathoms) and have sizes varying from a reef about 800 m. (0.5 mile) square to a circular one having a diameter of 11 to 16 km. (7 to 10 miles). The bottoms are variously described as consisting of rock, limerock, coral rock, coral, shells, or sand, and all are said to bear gorgonian corals and sponges. Under these conditions we can be sure that alge also occur there. To the north of Beaufort, rocks are said to occur off the coast of Virginia, and fishing grounds with rocky or sandy bottoms are listed by Goode and associates (1887, pp. 46-51, charts 12-13) off the coasts of Delaware, New Jersey, and the south shore of Long Island, connecting with those off the coast of New England. These conditions—the existence of a fairly continuous line of submerged rocky reefs extending from subtropical to cold northern waters, the subtropical nature of the flora found on the reefs offshore from Beaufort, the apparent existence of algze and corals on the bottom in the Gulf Stream, together with the northern course of this stream—seem to furnish excellent means for subtropical species of plants and animals to travel up our coast. Such species can live offshore in water warmed by the Gulf Stream, and, if the local conditions permit, may establish themselves temporarily or permanently on the mainland. These facts probably explain the occurrence of several of the species found in Beaufort Harbor and probably account for all the species found on the beach. It would be interesting to discover how far north of Beaufort subtropical species may occur. A thorough survey of these reefs, including their geology, oceanographical conditions, flora and fauna, would undoubtedly yield facts of great interest and importance. GENERAL ACCOUNT OF THE ALG. The intermediate location of this region gives a flora of considerable interest, containing both northern and southern elements, with southern species predominating. Considering the flora as a whole, of the 142 recognizable species and varieties recorded, 133 have been obtained in proper condition and amount for determination. Of this number, 62 (46.6 per cent) are found in New England, and 91 (68.4 per cent) are known to occur in the Florida-West Indies region. In the different divisions the relative numbers are as follows: Species and varieties of alge identified in Beaufort region. Division. Recorded for— Total. New England. | Florida-West Indies. Number. |Per cent.2| Number. |Per cent.b| Number. |Per cent. Myxophyceze To 7-5 7 70-0 8 Bo.0 Chlorophycez. .. a5 18.8 12 48.0 aI 84.0 Pheophyceze.... : 27 20.3 12 44-4 13 48.2 EEIRMIDARG RRR E Se ieeea ares eat ece esterase recite herr eters otare 7 Cy 3r 43-7 49 69.0 @ Per cent of total number identified in Beaufort region. > Per cent of total number in the division identified in Beaufort region. 110307°—21——25 380 BULLETIN OF THE BUREAU OF FISHERIES. Of the 84 genera found in the Beaufort region, 24 genera and 46 species reach here their northern known limit on our coast (Tables 5, 7), while 4 genera and 9 species reach their southern known limit in this region (Tables 6, 7). Furthermore, 20 species not previously recorded for North America have been found, 11 of these being new. Of the 133 identified species and varieties, 78 (58.6 per cent) are recorded for Europe, and 41 (30.8 per cent) for the Pacific coast of North America. The 46 species reaching their northern limit here (Table 5) have been found as follows: Growitig in BeauforthHarbors pera eile a ctie el rme cio oto ererotayaiete ls earns ete eater 16 Growingionly‘ott coral reef. 0720. 22 Oi REE Sees tees Alt eee 16 Found only onzBorte Beach hx) ui) 275 BRIGG. Bynes SOS iets oot ete II Known only,from other localities.....c-alatt -try-aemewey alts sos - 55 gat ae aaale 3 The 9 species reaching their southern limit here (Table 6) have been found as follows: Growingiit BeattiontrHarbOre. am cae ore ss ets teaeeiate vite ec nferteebelaissayiemiere: coe yG Growing only oncordl:secf!? . ROI.) Ey, lee, Re. Se. Bae I Known only from otherlocalities.ji.ic<.). , (aisehlon- bth k Qoit Wk SS SORE Seis B I The 20 species which are new to North America (Table 7) have been found as follows: Crowiiip ist BeditortsHarbors ask. nocache toe tin laseentetetcle or tncle ciertie eel et eaerel ae 6 Growingiqnlyjoncoralirecisz..2J ai. 29ST 9285 §. .,. IRS Se, SAS 12 Found only on Bopue Beach. -tetgilesies dates enced. Soasareute h-bieotes I Knowntonly tromiother localitiess. acts iam e-ytoteete cietetelesr-ire erotic eis oe oe ees I The most striking characteristic of the flora is the preponderance of red and the paucity of blue-green alge. The large number of red alge indicates the southern relationship of the flora; but here also is found a large northern element, as was shown above. The small number of blue-green alge is not easily explained. At other places the number is probably greater than is indicated here; indeed, the author saw large masses of undetermined Myxophycee covering the rocks of a jetty near Georgetown, S.C. At Ocracoke, N. C., also there were observed masses of blue-green algae densely covering the ocean beach just beyond the high-tide line for many square meters and covering the wharf piles between tide lines. The number of species found in these places was not large, but other species may have been present. At Beaufort, N. C., however, although one species (Lyngbya confervoides) is very abundant, covering walls and jetties for considerable areas between tide lines, repeated careful searches have failed to discover any other species in abundance and have yielded a total of only five species growing in the harbor. The relative richness, in other respects, of the Beaufort flora as compared with the flora of other localities is shown by the fact that of the 142 species and varieties recorded for the region 132 were found at this place. While a part of this numerical preponder- ance is undoubtedly due to the fact that Beaufort has been studied more thoroughly than other localities, a large part is due to an actually greater richness of the flora of this region. At no other locality has the author found anything to approach the number of individuals or of species that may be observed at Beaufort on a single collecting trip at any time during the summer. MARINE ALGA) OF BEAUFORT, N. C. 381 The 124 identified species and varieties recorded for Beaufort have been found as follows: Number. | Per cent. era watt ratie reser OE (GE ale wy aetiete «= Heteiciare es ctelelerale e|« = = bicisis,aidieis Son te sta inrshilo nis bs elele’a gig siniale.s theiotm binin’s idiots a's pin 77 62.1 SEO WATE Osh y OM ENE COCA VECIS (ABLE 2) rccist create vatele/atolerelcistetel startet clsteialctolaperatitatatet efa'ete's elsia,sietalebaralsynve/arniate\sia aiece/eis'p 29 23-4 ROcetirratl oustal Oe AOR MEA HCACK CLADIC SG) | rare ohrisleateisleloiatelsfelotalatcleteteteteleteteletete sinteisiaceia sists tecraieralecsialersleiaintetela sisiniate 18 14-5 As with other plants, two factors determine the algal flora of any region. First, the conditions prevailing at any place naturally exclude all species which are not able to grow under those conditions; second, of the species which are able to grow in any locality, only a part find access to the region and arrive there under conditions favorable for obtaining lodgment. We may be certain that there are hundreds of other species that could grow at Beaufort if they should be carried there. Since, with marine alge, artificial means of transport are usually excluded, the flora which we find in any locality favorable for the growth of alge is determined to a considerable extent by the direction of the currents bringing fruiting plants, fragments, or spores of algae from other regions. Occasionally, however, an alga may be introduced into a regicn by artificial means. On one occasion there was found in Beaufort Harbor a fragment of Halimeda sp. ‘This seemed a very interesting discovery until it was noticed that there was in the harbor at that time a boat from the West Indies bearing tropical-shells and other marine objects for sale. To this boat we may confidently ascribe the presence of the Halimeda. Although this species of alga did not establish itself at Beaufort, its presence there showed the possibility of the distribution of alge by artificial means of transport. There is evidence that at least one species has established itself at Beaufort during the progress of these studies. Rosenvingea orientalis, known elsewhere in North America only from Guadaloupe and from Wrightsville Beach, N. C., was first found on Bogue Beach in September, 1905, and was not observed in the harbor during that year. The following summer, however, this species was found growing between Fort Macon jetties and on the sea buoy, and in the summer of 1907 it was found on Shackleford jetty as well as on Fort Macon jetties. The records indicate similar facts for a few other species, but are not sufficiently complete to warrant conclusions about them. Miss Dunn (1917) has presented convincing evidence showing that one species of alge, Dumontia filiformis, appeared on the coast of Maine and established itself there between the years 1909 and 1913. ‘This species seems now to have spread in considerable abundance along a large part of the New England coast. Several species have been found growing in Beaufort Harbor on only one occasion. Such species, while obtaining a foothold, seemed unable to maintain themselves, perhaps because of changing conditions. These may be expected to reappear at any time and may establish themselves. Other species have been found only occasionally, being represented by scattered individuals. Such species seem to be living near the limit of their endurance and may appear and disappear as conditions become more or less favor- able. Still other species, not yet observed here, may be expected to appear whenever chance currents bring them to this region under conditions favorable for their obtaining a foothold. 382 BULLETIN OF THE BUREAU OF FISHERIES. FLORA OF BEAUFORT HARBOR. Considering, first, the 77 species and varieties found in Beaufort Harbor, the number of these in the different divisions is: Number.| Per cent. Myxophycez.. . 5 6.5 Chlorophycez. . 17 22.1 Phzophycez... daa 15 19.5 Bh odo phi y ce sings: wud ewinatecitereeenn va cure eee Ree Te wee ele ee ee ee eee Bee ene een ne 40 51-9 These are distributed throughout the year as follows: Statues flora onlyjg Fait ro= TLR. TE a Savi Gabor coe th cave oe fait. Seu, PMEN REE ake alan aanien ace 4° 51-9 Spring flora only......... 22 28.6 Spring and summer floras 4 5-2 Perennial Ir 14-3 The strictly summer flora is distinctly southern in its character, but even this has a decided northern element. Of the 40 species and varieties included here, 30 (75 per cent) occur in the Florida-West Indies region, while 17 (42.5 per cent) are found in New England. Of these 17 forms recorded for New England, however, all except four are of general distribution, occurring in the Florida-West Indies region also. The distribution of this summer flora in the different divisions is as follows: Number.| Per cent. MY ODA YOR os ri ispecies «+ Hob op Rieio pet = = ben lobleeetall- « Deedsat ee se sto heid= eb ppb «de tetge- «att: Soe ke keer 4 10.0 Chlorophycez. . 5 7 175 Phzophycee... waa Er, 7 17-5 ROMO PH Vent aces cost acnaais cinainre: Unie scatnie me PARE ota tia state tah Te aia cimialaseiafersisiclaicelele Scemaiste else ale ate te caters 22 55:0 The strictly spring flora, on the contrary, is distinctly northern, of the 22 species and varieties recorded, 20 (90.9 per cent) being found in New England and only eight (36.4 per cent) being known from the Florida-West Indies region. Its northern character is further shown by the fact that red alge do not predominate here, the number in the different divisions being: Number.| Per cent. Bi yop hy cea tose 55 spare orapaie sere hates Wie fae cio etwas ein taro als! loyal evar ec Bio, ow bin Nin alee Re ie At ae er ° 0.0 Chlorophycez. . 8 36.4 Phzophycez... 6 27.2 Rhodophycez.. 8 36-4 The four species common to the spring and summer flora are red alge. All of these are found in the Florida-West Indies region, while three occur in New England also. Of the 11 perennial species, all are found in New England while nine occur in the Florida-West Indies region also. The numbers in the different divisions are: Number.| Per cent. Dy a Tr 0) eho OBESE CIEE RC eA eI eae citi Ca O54 | AA ABE RODE cE SSE AIS Ue SIIEIMECIADS SEpICIaarTS is I 9-10 Chlorophycez. .. . 2 18.18 Phzophycez..... a 18.18 Rhodophiycéas. ssi a8 Fe cheek ila are sand. dee tae eked aE Picla cera cotbine Ske. Lauekiicte opisite cetke b 6 54-54 It is probable that further search would increase the number of species in this list. _ alll MARINE ALGA) OF BEAUFORT, N. C. 383 FLORA OF CORAL REEFS. The flora of the coral reefs is predominantly southern, of the 47 identified species and varieties found there (Table 2), 32 (68.1 per cent) being recorded for the Florida- West Indies region and 14 (29.8 per cent) being known from New England. Comparing the three collections made on the principal reef, we find the species occurring as follows: Species of algz identified for coral reefs. Date collected. Recorded for— New England. |Florida-West Indies. eee eee eee Number. | Number.| Per cent.| Number.| Per cent. 8s. PN iy ernie ee ey ata cote kia niet wabotecnin a ieieiele\n\ ain Wve oleate, slaibselas claie arsie'sia 21 9 42-8 18 5-7 PATER USE OES. A Parad 2 Metcatin o/c NARA. «aisle Bin ad etleiccehlnade he bioktctaaat shale «nnd as 6 24-0 15 60.0 A Sue STs Se Gas ego JOIN ODOC OOD IDOARD ES UCACh EE ae erPEOeEmnat aa 10 45°4 20 90-9 This southern relationship is more striking when it is remembered that the visit to the reef in May was made at a time when Beaufort Harbor bore the spring flora, having 90.5 per cent of the species common to New England and only 33.3 per cent of the species common to the Florida-West Indies region. At this time several northern species which occur in this locality only in the spring were found on the reef. The small proportion of the species collected in August, 1914, which are common to other regions is due to the fact that four of these are new, while six are new to North America. If these species are excluded, the figures are New England 4o per cent, Florida-West Indies 100 per cent. Similarly, if two species new to North America collected in July and August, 1915, are excluded, the figures for this period are New England 50 per cent, Florida-West Indies 100 per cent. It could not be illustrated more forcibly that this flora is southern in its nature and that the species which are common to New England are those which are generally distributed and occur along the entire coast. On all trips there were obtained from this reef species which were not found growing elsewhere in this region. In May, 1907, ten such species were observed, eight being distinctly southern, one being distinctly northern, and one being generally distributed. In August, 1914, there were observed, besides the ten species that are new or new to North America, eight species not found elsewhere in this region, seven being distinctly southern and one being generally distributed. In July and August, 1915, there were collected, besides the two species that are new to North America, nine species not found elsewhere in this region, seven being distinctly southern and two being generally distributed. It is thus seen that the flora that in this region is confined to the reefs is overwhelmingly southern in its relationship. Only four identi- fied species were found in all three collections from the reef, while 14 were obtained in two collections, and 29 were found only once. This is probably an indication not so much of a seasonal distribution as of the abundance of the species occurring there and our ignorance of them. It is highly desirable that a thorough study of these reefs be made. 384 BULLETIN OF THE BUREAU OF FISHERIES. The southern character of the species occurring on this reef is further shown by the predominance of red alge. The total numbers found in the different divisions are: Number.| Per cent. My xophy cena, oi ois. Foleo Bee Te eee eee oka ale aa ae a Bria. ae eae SCR Ret INT Wak cial ate tehie cn Peds aie 2 3-8 Chlorophycez PE 4 75 TREMOR COINS ois Sic sata cs arate a fo cl aratatujoasralefereba ye tace eyeiminie el Siele WiSP ae ieyele, arcade slajeiaveysceretsi het efalatale e(hysteneter «fete 544 Ir 20.8 Rhodophycez 36 67-9 FLORA OF BOGUE BEACH. The flora which, in this locality, is found only on Bogue Beach (Table 3) is as pronouncedly southern as is that of the coral reef. Of the 18 identified species and varieties composing this list, 16 (88.8 per cent) are known from the Florida-West Indies region, while only four (22.2 per cent) are recorded for New England. ‘This relationship is again shown by the predominance of red alge. The total numbers found in the different divisions are: { Number.| Per cent. Meyxophiyceses§ forfoo.. etter aoe See eee eke o UE oteld Borda ee Alldald Week teleed ate alialdaMatieleeihtkis Pee omonae haa I 4.6 Chlorophycee. . | 4 13-6 Pheophycee...... 7 31-8 Rhodophyceze Ir 50.0 These facts support the suggestion previously made that most of these specimens have been washed in from the coral reefs offshore or from the reefs lying to the south of this locality, while some may have been brought by the Gulf Stream from the Florida- West Indies region. CONDITIONS AT BEAUFORT, N. C. HARBOR. The principal factors affecting the growth and distribution of alge are temperature, light, composition of the water, turbidity, movement of the water (including tidal range), and the nature of the habitat. TEMPERATURE. The temperature of the surface water at the laboratory wharf (on Pivers Island) has been taken at 5 p. m. almost daily during three periods, totaling almost four years. A full statement of these figures is given in Table 9. A summary of the records, stated in degrees centigrade, is as follows: Change of Change of Maxi-) Mini-] Aver-| 9¥te8° Maxi-| Mini-| Aver-| “yirege mum.|mum.| age. Rae viol mum.|mum.| age. | previous month. month. tf SG: FG: a ee eG "GC, T5.5 5-0 9-8 31-0] 25-0] 27-9 +2.7 16. 7 3-0 9-6 30-0] 23-0] 27-5 — 4 19-5 3-0] 12.5 28.9] 17-8] 24.8 =e 23-0 12-0 17-5 25-0 14-0 19-2 —5. 26.7 18.0 22.4 20.0 7-8] 13-9 —5-3 1 (oes paascstonr ceca: 2. go-0}] 17-0] 25.2 17-0 6.0] 11-1 —2.8 MARINE ALG OF BEAUFORT, N. C. 385 It will be seen that the extreme range of temperature recorded is 28°, from 3 to 31°. The lowest temperature and the lowest average occur in February, while the highest temperature and the highest average occur in July. In the fourth column there is given the change of the average since the previous month. It will be observed that the greatest increase of the average occurs from March to April (5°), while that from April to May is only 0.1° less (4.9°). The greatest decrease of the average occurs from Sep- tember to October (5.6°), while that from October to November is nearly as great (5.3°). During the other eight months the average change is relatively small. It is interesting to compare with these figures the surface temperatures (expressed in degrees centigrade) recorded for Woods Hole, Mass., by Sumner, Osburn, Cole, and Davis (1913) and those given for Naples by Berthold (1882), since the former locality has a temperate algal flora, while the latter locality has a subtropical one. Woods Hole, Mass. Woods Hole, Mass. Naples, SITPRO SGT OID cada lt Naples; Maxi- | Mini- | Aver- Italy. Maxi- | Mini- | Aver- Italy. mum. | mum.| age. mum. | mum, | age. rd OF, 5G. °C: of CP bet Coa apotarmciy scence 23-6 17-22 | 20.97 25-27 21-67 | 17.22] 19-55 } 18.33 Io. 83 15-26 12. 78 3-6 Sot teas 8.6 — .28 Cm nwt tome artes At both of these places the highest temperature occurs in August and the lowest temperature in February. Woods Hole has a range of 25.55°, from —1.95 to 23.6°, while Naples has a range of 19°, from 8 to 27°. These figures indicate that Beaufort has a higher maximum and a lower minimum than Naples; but the record of Naples is less complete than that of Beaufort. LIGHT. While we have as yet no satisfactory measure of light, we can measure, in an approxi- mate way, the relative effect under different conditions of the rays of light which affect photographic paper. This has been done in the present instance by means of the Clements photometer. This instrument uses a strip of solio paper, successive portions of which are exposed at will through a small slot, the slot being opened or closed as desired by means of a sliding cover. Standards for comparison are obtained by exposing portions of the paper to direct sunlight for different measured intervals of time. Another portion of the paper is exposed for a definite time in the situation whose light is to be tested. By comparison it is then determined which of the standards is darkened to the same extent as the paper exposed in the test situation. From the relative time of exposure of the test paper and this standard it is thus possible to estimate the relative amount of light in the test situation compared with full sunlight. For example, if paper exposed in a certain situation for 10 seconds is darkened to the same extent as a standard exposed to full sunlight for 5 seconds, we estimate that the light in this situation is 50 per cent as strong as full sunlight. It is, of course, necessary to make new standards for every series of tests, since the intensity of full sunlight will itself vary at different times and on different days. 386 BULLETIN OF THE BUREAU OF FISHERIES. Since the only photometer available to the author at that time was one intended for use in the air, this form was employed, being adapted as follows: All exposures were made with the photometer placed in a glass preserve jar of sufficient diameter to permit the photometer to lie flat on the bottom. The photometer was held in place by paper packed into the jar, care being taken that the slot for exposing the solio paper was not shaded by the packing. Ina dim room the slot was opened, the photometer was placed in the jar and securely packed, and the jar was tightly wrapped in black cloth. This was then taken to the desired situation, the jar being held horizontally with the slot directly on top, the cloth was quickly removed for the desired number of seconds and then quickly replaced, and the jar was then brought back into the laboratory. All changes in the apparatus were made in a dim room at a considerable distance from any window. The standards were obtained in this way by exposing the photometer within the glass jar to direct sunlight on an upper, unshaded, southern porch. For exposing the photometer below the surface of the water a shallow box, open at the top and of the proper size to hold the jar horizontally, was built, the sides of the box being just high enough to hold the jar in place and not shading the upper part of it. This box was then fastened to a handle marked with the desired distances. In a dim room the jar containing the opened photometer was placed horizontally in this box with the slot directly on top, and the box was tightly wrapped with black cloth. This was carried in a boat to the desired locality and held at arm’s length below the water, the black cloth was then removed, and the jar immediately sunk to the desired depth and held at that level for a definite time. The jar was then quickly brought within reach and imme- diately covered with the black cloth, not more than a second being required for this manipulation. The apparatus was then carried to the laboratory, where all changes of the photometer were made in a dim room. While the jar undoubtedly diminished the light reaching the photometer, this decrease would be the same in the standards and the tests. The effects of the light in these two cases may, therefore, be directly compared. Two records, one at high and one at low tide, were obtained in this way in the channel in front of the laboratory wharf in July, 1907. In the first of these the standards were made from 1:15 p. m. to 2 p. m., July 17, and the measurements below the water were made from 1 : 15 p. m. to 4 p. m., July 18, high tide on this day occurring at 2 p.m. In some cases, where the color of the test did not exactly match that of any standard, the time of the standard having an effect equivalent to that of the test was obtained by interpolating between the two standards showing the colors nearest to that of the test. Standards were made by exposing to direct sunlight as described above for 60, 30, 25, 22, 20, 15, 10, 5, 3, 2, and 1 second. ‘The results were as follows, the first column giving the depth below the surface at which the test was exposed, the second column giving the time of the exposure of the test, the third column giving the time of exposure of the standard having a color equivalent to that MARINE ALG OF BEAUFORT, N. C. 387 of the test, and the fourth column giving the calculated percentage intensity of the light at the respective depths compared with full sunlight: Depth Length of | Equivalent] Relative EISEN) exposure. | standard. | intensity. Seconds. Seconds. Per cent. 6 Te) 28.0 46. 60 23-0 38-3 60 22.0 36.6 60 9-0 15-0 60 5-0 8.3 120 7-0 5-8 180 4:55 2-5 240 2.8 1-2 300 3-0 1-0 The second record was made from 1:15 p. m. to 2:30 p. m., July 24, low tide on this day occurring at 2:15 p.m. The standards were made at 2 :30 p. m. of the same day and were exposed for 60, 45, 30, 25, 20, 15, 10, 5, and 3 seconds. The results were as follows: Depth Length of | Equivalent} Relative wees exposure. | standard. | intensity. Seconds. Seconds. Per cent. 60 5 60 17-5 60 2 120 3 120 3: 120 3 » peeved MAMA A While these two records differ considerably, they agree in their main points and indicate several interesting conclusions: (1) A considerable portion of the light (nearly one-half) did not penetrate below the surface, probably because of the reflec- tion from the water and the suspended matter; (2) of the light which entered the water nearly one-half did not penetrate to a depth of 30 cm.; (3) at a slightly greater depth (1.2 m. at high tide, 60 cm. at low tide) the light was so reduced as to be almost lacking. These results are of great interest when considered in connection with the vertical distribution of the alge. While some of the difference in the records may be due to errors in the determinations, a considerable part is probably due to the fact that one was taken at high and the other at low tide. The water at high tide is notably clearer than that at low tide, and the record taken at high tide shows a correspondingly greater light intensity. These records, of course, show the effect of only the rays affecting solio paper, but it is these rays (toward the violet end of the spectrum) that are least absorbed by water. It is not known what proportion of the different rays penetrate water as turbid as that occurring here, or what is the intensity of the rays at the red end of the spectrum that reach the slight depths at which these measurements were made. There are evident errors in the methods used, but since the figures obtained could be, at best, only approximations, it did not seem worth while to give the time necessary to improv- ing the records. The figures given refer only to the water in the channel in front of the laboratory wharf. Efforts to obtain records from other localities far removed 388 BULLETIN OF THE BUREAU OF FISHERIES. from the laboratory were not successful, since the necessary changes of the photometer could not be made out of doors. These figures probably represent about an average of the conditions occurring through the greater part of the harbor. At certain places, especially near the inlet where the alge are more abundant, the water is somewhat clearer. SALT CONTENT OF WATER. Determinations of the salt content of the water from five places in Beaufort Harbor were made by Wheeler (1910) in the summer of 1909 during the progress of the present study of the alge. The water was obtained from (A) Beaufort Inlet; (B) the laboratory wharf; (C) Bogue Sound opposite Moorehead City; (D) between the eastern end of Beaufort and Bird Island Shoal; (E) Green Rock in Newport River near the entrance to Core Creek. The results, stated in parts per 1,000 grms. of water, were as follows: A. B. (Ge D. E. 28.043 27-836 27-977 28.006 24-796 842 +742 +751 +751 = 702 3-379 3-245 3-300 3-335 2.978 2-417 2.328 2.320 2-372 2.062 I. 171 1.168 1.202 1.188 1-039 -220 +214 +214 +215 +215 Wotal peeves te sets pigeie eth. thee ore SO Sete a OR Ce 36.072 35-533 35-764 35-867 31. 786 Specific gravity at 28.7°C.............. Be ope erica done rise 1.0227 I-0222 1.0226 1.0227 T- 0193 As is shown, both the total salt content and the relative amounts of the different salts vary in different places and at different times, the total ranging, in these analyses, from 3.1786 per cent to 3.6072 per cent. The density, of course, varies at different times, being largely determined by the amount of rain and the state of the tide. At times, after continued hard rains, the water in the harbor has, for days, the color of weak, muddy coffee, due to water coming from the inland juniper swamps. Daily salinometer readings have been made at 5 p. m. at the laboratory wharf (on Pivers Island) since June, 1913. A summary of these is as follows: Maximum.| Minimum. | Average. Maximum.| Minimum, | Averege. Jirme wa este ac ceetonsiet 1.0228 1.0184 I. 0209 1.0248 1. 0186 I- 0212 Te ie ak a Oe 1.0238 1.0216 1.0228 1.022 1-010 I-O179 AMGgHSES . FLAC. BE 1-0226 1.020 I. O21 1.0204 I. O1r2 I- 0173 September.......... 1.0204 t.0132 1.0168 1.0213 1-015 I- 0183 ctoher7s th SAGs 1.0236 I. O17 I. 0199 1.0234 I.019 I-0212 November........... Ae I.024 I. 0102 I.0209 1.0258 I.021 1.0234 December’, AGLI ase 1.0256 I. 0192 1.0226 1.0246 1.022 I-023 It will be observed that the recorded density ranged from 1:o10 to 1.0258. The general average, obtained by averaging the monthly averages, is 1.0205. In these figures no account is taken of the temperature, since in such salinometer records the errors of reading are almost certainly greater than the temperature corrections. For the same reason the maxima and minima are not accurate, but probably cover the MARINE ALGA) OF BEAUFORT, N. C. 389 range of variation. The averages, however, are probably fairly accurate, since they are obtained from a large number of readings where the errors probably balance each other. The general average, 1.0205, may, therefore, be taken as closely approximating the mean density of the water at the laboratory wharf. At other places in the harbor the density will, of course, be different from this. Since alge grow throughout the harbor, some of them will be exposed to greater densities and some to lesser densities than those recorded here. Several salinometer readings have been made by the author at other places in this region. While these have not the value of the daily records made at the laboratory wharf, they indicate the comparative density at other places. They are as follows: Newport River near “Green Rocky? low tide... ).04.. 10 MA 0 ohare ats I. o16 North ‘River near Lenoxville low tide. .otisoesh. onine CUA ek I. 0188 Pamlico Sdund,-Ocracoke low:tidess .vregigeinavc. oleleieciosscteas le madd. I. OIL Goralimeelolt Beautorts actpeh reppin cites tacit oi ean eumoias mys aeetaette corse icici I. 0242 TURBIDITY. The water from the open ocean outside of the inlet contains a considerable amount of suspended matter, as is evident when this water is filtered, while the water within the harbor has still more fine, suspended matter and is, at times, very turbid. All rocks, shells, and posts under water are soon covered with a thick deposit, and at many places in the harbor the bottom is covered with mud upto a meter or more in depth. In the harbor and in Bogue Sound the amount of suspended matter seems to increase as we go farther from the inlet, while in the sound back of Shackleford Banks the water is decidedly clearer, owing to the strong current running in from the ocean at this place. Farther back in this sound the water is as turbid as in the harbor. This turbidity not only reduces the light penetrating the water but itself affects the vertical distribution of aige, since much of the suspended matter is deposited on all objects in the water. ‘The older portions of the broader alge (as Dictyota, Padina, the leaves of Sargassum) are more or less thickly covered by this mud settling from the water. MOVEMENTS OF WATER. The usual maximum range of tide (at the spring tides) is 0.97 m. (3.2 feet), the usual minimum range (at the neap tides) is 0.7 m. (2.3 feet), and the usual mean range is 0.82 m. (2.7 feet). The tides may, however, vary considerably from these figures, the actual height and range attained depending in part on the direction and strength of the wind. The greatest range observed by the author at the laboratory wharf is 1.31 m. (4.3 feet). The smallest range observed is 0.48 m. (1.6 feet). Under excep- tional conditions the low tides are higher than the high tides recorded on other days in the same month, while at other times the tides are unusually low. Although there is not a very great difference in the height of water at high and low tides, there is a great difference in the amount of light reaching the alge at these times. Except on Shackleford jetties, where the water is clearer, no alge were found in the harbor below 1.4 m. below low water, and the majority were found within 75 cm. below low water. Most of the alge have, therefore, about twice as much water over them at high tide as at low tide. Furthermore, during summer and autumn the greater number of alge grow almost up to the surface of the water at low tide. For these parts of the plants 390 BULLETIN OF THE BUREAU OF FISHERIES. the difference in the amount of water covering them at high and low tides is much greater than is indicated by the figures given above. For the species growing above low water the difference is, of course, still greater. Since light penetrates to such a slight depth in this water, the difference in the amount of light received by the plants at different stages of the tide must be very great. The difference is, however, partly neutralized for the alge growing near the inlet and even as far back as the laboratory, since the water of the ocean is clearer than that of the harbor. This ocean water, entering the harbor at flood tide, pushes the more turbid water before it and mixes with it, so that, as was shown above, the water of the harbor is clearer at high than at ow tide. Since the harbor is a comparatively small body of water and is well sheltered by land, the water is usually smooth throughout the greater part of its area. Near the inlet, however, there is considerable movement, although even here there are usually no waves. Even the slight movement that does occur here, however, probably affects the algee growing on Fort Macon and Shackleford jetties by washing off the sediment that settles on them. HABITATS. The bottom throughout the harbor consists principally of sand, with some areas covered by mud or shells. (See map 2.) The mud and sand furnish no place of attachment for alge. ‘The shells furnish excellent places for attachment, but do not bear alge, probably because of the turbidity of the water above them. Algz are, however, found attached to single shells and other supports below low water along the shore and, sparingly, on the shoals. The numerous wharf piles occurring here would seem to offer excellent habitats, but during the summer and autumn no alge were ever found on them, while, on the contrary, during the spring they bore an abundant growth of alge. The reason for this difference is not apparent. There remain, as possible algal habitats, the jetties at Fort Macon, at Shackleford and on the laboratory island, and the brick walls occurring along the town shore. These jetties bear the greater number, both of species and of individuals, of the alge growing in the harbor, while the walls bear a limited number of species. Small species of alge are found, too, in some abundance on the buoys that mark the channel into the harbor. CORAL REEF. The physical conditions existing on the coral reef offshore have already been described. Here it need be said only that the surface temperature at noon on May 15, 1907, was 21.11° C., the temperature at a depth of 25.5 m. was 19.44° C., and the density of the surface water (measured by a salinometer) was 1.0242. At this time the tempera- ture of the surface water in the harbor varied, in different places and on different days, from 20.5° C. to 23° C., and the density varied from 1.0165 to 1.0212. — MARINE ALG OF BEAUFORT, N. C. 391 DISTRIBUTION OF ALGA AT BEAUFORT. REGIONAL. The regional distribution of alge, their occurrence throughout the world, is, like that of other plants, determined largely by temperature. Whether a given algal species is able to exist in a given locality will depend absolutely on its ability to endure the maximum and minimum water temperatures occurring in that locality. It need not, however, be obliged to endure these extreme temperatures in its vegetative condition, but may exist for long periods by means of spores or fragments, resuming its vegetative state with the return of more favorable temperatures. Setchell (1915) has shown that the majority of the species of alge occur in regions having a range of not more than 10°C., and that those occurring in regions having a greater range than this accommodate them- selves to the general law by their seasonal distribution, etc. Of the species which are able to exist in any given locality, some will thrive and will predominate, others will barely maintain a foothold, while others will appear and disappear at different times. The relative abundance of the different species occurring in any locality will be determined by the ability of these species to thrive under the conditions found there and to compete under these conditions with the other species growing in the region. To become abundant, a species must be able not only to endure the extreme conditions, but also to grow luxuriantly under the usual conditions. The factors most affecting the relative abundance of the species of marine algee seem to be the temperature, density, and turbidity of the water, and the intensity of the light occurring, not on single days, but throughout the growing season. As has been mentioned, the intermediate position of Beaufort makes its flora particularly interesting. Here Codiuwm tomentosum, Dictyota dichotoma, Padina vickersia, and other strictly southern forms grow along with the more northern Fucus vesiculosus and Polysiphonia harveyi. As a rule, however, the northern and southern species do not grow together, the former occurring in the spring and the latter in summer. Setchell (1915), in considering the effect of temperature on the distribution of alge, distinguishes the following regions, based on the average temperature of the water during the summer expressed in degrees centigrade: Lipperisboreal sp aicsi- Fyn shrewd oraryetia- 4 o to 10 | South subtropical. ..............0000000s 20 to 25 Lower horeal ss. 26/02. 50 |. gras tesa s+ To to.r5 | \So1thtemperates vo ose. csiers sisjasogte =~ 15 to 20 North temperate yet core cies og cise vis eke De tO zon LOWER AUSIR A cc mace ec eaericleltoemay ond eer to to 15 INGIiLSMULOpICdl ener arc et crea ices BOWOTAG | MUPPEL AUS AL yc cents oo eaten ee isteisiete ts o to Io Tropical’; 212 SIQEL i. Dah. ee 25 N.to25S. Since the average water temperature at Beaufort during the months from June to September is 26.35 °C., this classification would place the Beaufort area at the northern limit of the tropical region. It would seem, however, that the Beaufort flora should more properly be included in the subtropical region, and the limits given by Setchell should be modified. SEASONAL. As is shown in Table 1, the flora of spring and summer are very different. Of the 77 species and varieties growing in the harbor, only 15 (19.5 per cent) are found at both seasons, 11 of these being perennial and 4 having been found from April to October. 392 BULLETIN OF THE BUREAU OF FISHERIES. During the winter of 1908-9 monthly collections of all species observed were made for the author from the time of his departure from the laboratory, October 20, 1908, until his return, June 30, 1909. While more careful studies would probably alter the present data in some details, these collections and those made by the author in May, 1907, and April, 1908, give a fair picture of the seasonal distribution of the alge in the harbor. The species occurring there in the summer may be grouped as follows: DOMINANT SPECIES. Erythrocladia recondita. ; Erythrotrichia carnea. Lyngbya confervoides. Goniotrichum alsidii. Codium decorticatum. $ Acrochetium dufourii. Codium tomentosum. etek nether Sargassum filipendula. Acrochetium virgatulum. Boas CSE Gelidium coerulescens. seer ie te Gelidium crinale. ap ees, Sigua apa Gymnogongrus griffithsiz. SOE pon mite tee Actinococcus aggregatus. a eats ae Hm Agardhiella tenera. wreepa ie S Eucheuma gelidium. Dermatolithon pustulatum. Chatapia waryila OCCURRENCE MORE LIMITED. Lomentaria uncinata. Chondria sedifolia. Ulva lactuca. Herposiphonia tenella. Ulva fasciata. Polysiphonia harveyi. Ectocarpus mitchelle. Polysiphonia denudata. Rosenvingea orientalis. Ceramium tenuissimum. Fucus vesiculosus. The other species of the summer flora mentioned in Table 1 have been found only occasionally. By the middle of October changes in the flora have become evident. Dictyota has become relatively scarce, and Padina is less abundant than formerly, the plants of both species being small and showing signs of degeneration; Chondria dasyphylla has almost disappeared; Hypnea is still abundant and shows little change except that there seems to be a greater preponderance of tetrasporic plants than formerly; Codium and Gra- cilavia confervoides are still present; Ectocarpus is abundant and conspicuous; large plants of Fucus are abundant and the fruits are well developed. During November this change continues. Dictyota becomes still scarcer and finally disappears; Hypnea is reduced to small sterile plants, the condition in which it passes the winter (PI. CI, fig. 2); and the first of the spring flora, Petalonia fascia, makes its appearance. In December we find the last plants of Padina and the species growing on this, Erythrotrichia and Goniotrichum; Gracilaria confervoides has disappeared; Codiwm tomentosum, Champia, and Chondria sedifolia are still present; Enteromorpha prolifera, Agardhiella, and Gracilaria multipartita are more conspicuous; Petalonia grows to a large size; while occasional plants of Grinnellia and Dasya are found. By January Champia and Chondria sedifolia have disappeared; Ectocarpus conjer- voides has replaced the summer species, E. mitchelle; Grinnellia has become conspicuous, and small plants of Porphyra have appeared. MARINE ALG OF BEAUFORT, N. C. 393 In February we find the last reduced plants of Codium tomentosum (C. decorticatum having disappeared earlier) and of Chondria dasyphylla. From this time the flora consists entirely of perennial and spring species. Porphyra has attained a large size; Enteromorpha linza has appeared; and Enteromorpha prolifera, Ectocarpus confervoides, Petalonia, Porphyra, Agardhiella, Gracilaria multipartita, and Grinnellia are the principal species composing the flora. In March, the alge are scarcer than at any other time during the year, but the perennial and spring species already mentioned are present without change, except that small plants of Leathesia have made their appearance. During April the spring flora attains its greatest development. The dominant form throughout the harbor and along all the shores is Ulva lactuca, which occurs on all rocks and forms large masses lying free on the bottom. Closely rivalling this are Entero- morpha prolifera and Porphyra growing on all rocks and posts throughout the harbor. These three species are extremely abundant everywhere, but in limited areas they are surpassed by Polysiphonia nigrescens and Ceramium strictum. ‘The species occurring here at this time may be grouped as follows: ABUNDANT. Polysiphonia nigrescens. Dasya pedicellata. Lyngbya confervoides. Ceramium strictum. . Enteromorpha linza. Enteromorpha prolifera. OCCURRENCE LIMITED. Ulva lactuca. Ectocarpus confervoides. Enteromorpha flextosa. Petalonia fascia. Enteromorpha intestinalis. Myrionema strangulans. Chzetomorpha melagonium f. rupincola. Leathesia difformis. Rhizoclonium riparium. Fucus vesiculosus. Cladophora flexuosa. Porphyra leucosticta. Bryopsis plumosa. Acrochetium virgatulum. Ectocarpus siliculosus. Gelidium ccerulescens. Stilophora rhizodes. Agardhieila tenera. Sargassum filipendula. Gracilaria multipartita. Bangia fusco-purpurea. Champia parvula. Acrochetium corymbiferum. Lomentaria uncinata. Gelidium crinale. Grinnellia americana. Gymmnogongrus griffithsiz. Chondria tenuissima var. baileyana. Hypnea musciformis. Young plants of several species were observed at this time. Several specimens of Codium 3 to 12 mm. tall were found on shells in the clearer, deeper water north of the laboratory. Fucus showed, in addition to the large plants, many germlings 2 to 3 mm. tall. Small plants of Chondria dasyphylla also were observed. Besides the germlings of Fucus, many large plants up to 14 cm. tall were present. These were entirely without fruit. Ulva, Enteromorpha prolifera, Lomentaria, and Champia grew more abundantly and to a larger size than in summer. Fruiting plants of Hypnea 1 to 6 cm. tall were observed, all of these being tetrasporic. During May the spring species begin to disappear, some of the plants showing signs of disintegration. Enteromorpha linza, E. prolifera, Chetomorpha melagonium jf. rupin- cola, Stilophora, Bangia, Porphyra, Dasya, Polysiphonia nigrescens, Grinnellia, and Ceramium are, however, still present. Hypnea has now attained its summer condition, 394 BULLETIN OF THE BUREAU OF FISHERIES. the plants reaching a size of 22 cm.; Chondria sedifolia has appeared, and one small mass of Rhodymenia palmetta was found on Fort Macon jetty. By June the spring flora has disappeared and the summer flora is established. The growth of the summer species at this time is very rapid. On May 14, 1909, the jetties at Fort Macon were carefully searched for Dictyota without revealing a trace of this species. On June 9, when the next collection was made, there were found numerous plants 20 to 29 em. tall which had matured and liberated their sexual cells. The species present now include well-developed fruiting plants of Codium, Dictyota, Padina, Hypnea, Chondria dasyphylia, C. sedifolia, and Herposiphonia, and plants of Rhodymenia 6 cm. tall. None of the spring species was collected at this time. All of the summer species are present before the first of July and maintain them- selves until the following October or November. From these facts we can picture the seasonal succession as follows: With the advent of colder temperature, the summer flora begins to disappear by the middle of October, the larger number of the species disappearing by November or December, others dropping out with each successive month, but some remaining until February. The first of the spring flora makes its appearance in November, other species appearing with each successive month, the flora, however, remaining relatively sparse during the winter, the smallest number of species being found in March. With the coming of warmer tempera- ture, this flora becomes more abundant and reaches its greatest profusion in April, after which time it begins to dwindle and disappears by June. ‘The first of the summer flora appears in April, others appear in May, and all are present before the last of June. If the seasonal behavior of the alge is compared with the recorded water tempera- ture, it is observed that the disappearance of the summer flora in October and November is coincident with the greatest decrease in temperature; the appearance of the spring flora in November and the succeeding months follows this diminution; the time of greatest scarcity of alge, in March, follows the lowest minimum temperature reached; the rapid increase of the spring flora to its maximum in April is coincident with the greatest increase in temperature, while its disappearance during May is coincident with the continued increase; and the appearance of the summer flora in April is coincident with this greatest increase and its profusion in June follows this great increase of tempera- ture. It would scarcely be possible to find a more direct relation between temperature and the seasonal distribution of plants than is shown here. From this it seems evident that, while light probably has its effect, the seasonal distribution of alge is determined to a very great extent by the temperature. More exact studies would probably show interesting relations between the temperature and the individual species occurring here. The manner in which the summer species exist during the winter and the spring species exist during the summer at this place has not been determined. During the seven summers spent at the Beaufort laboratory, two small plants of Grinnellia and a few small plants of Dasya have been observed, but no other of the spring species has been found here after May, and none of the summer species has been found after February. Lewis (1914) has shown that, at Woods Hole, Mass., many of the summer species of red alge occurring there (Dasya, Polysiphonia, and others) persist during the winter by means of the minute holdfasts of sporelings, the other portions of these sporelings and all of the older plants dying at the approach of cold weather. The plants arising from these holdfasts the following summer were mainly tetrasporic. Probably some such method _ MARINE ALGA OF BEAUFORT, N. C. 395 carried the species over the unfavorable seasons at Beaufort. The rocks on which Dictyota and Padina grew the preceding summer and on which they occurred abund- antly the following summer were carefully searched by the author under favorable conditions in April, 1908, without revealing a trace of these species. It is probable, however, that a microscopic examination would show these and other alge present on the rocks below low water. It is interesting to note that, although cystocarpic and tetrasporic plants of Hypnea are present in the summer in about equal numbers, a collection of 55 plants of this species taken at random in October showed 45 tetrasporic plants and ro sexual ones, and all the fruiting plants observed in April were tetrasporic. Lewis (1914) has shown that the preponderance of tetrasporic plants in the early summer exhibited by the annual red alge at Woods Hole is due to the fact that the two generations are produced alternately, the last crop of the summer being prevailingly sexual, and the carpospores borne by this crop producing the sporelings whose holdfasts persist through the winter. The peren- nial algze at Woods Hole show no such discrepancy in the numbers of sexual and tetra- sporic plants. In the present instance it seems that the tetrasporic plants of Hypnea, a perennial species, are themselves more resistant to cold than the sexual plants. Fur- ther studies are needed on this subject both here and in other regions. The seasonal life cycle of Fucus may be summed up here for comparison with other regions. Young plants were observed in April along with large, old, sterile plants. The swollen receptacles become evident about June, but remain small and inconspicuous during July, becoming gradually larger and more conspicuous during August and Sep- tember, and reaching full size about the latter part of October, the plants showing large, well-developed fruits from November to January or February. After this time all plants observed were sterile. It is of interest to note that in May, 1907, when Beaufort Harbor bore almost entirely a spring flora, the coral reef offshore bore such strictly southern forms as Udotea, Dictyota, Zonaria, Nitophyllum, Chrysymenia, and others, along with the spring species of Dasya and Grinnellia, although at this time the water at the depth of this reef was at a lower temperature than that in the harbor. The explanation of this can not be given surely without further study, but certain differences between the harbor and the reef are evident. The greater clearness and higher salinity of the water over the reef probably play a part, but the chief factor probably is that the water at the depth of the reef, as may confidently be believed, does not fall to the low temperature found in the harbor in winter. This suggestion is supported by the species found on Bogue Beach during the winter from December, 1908, to March, 1909. Besides the species growing in the harbor, there were found during this time Zonaria flava, Z. variegata, Nitophyllum medium, Polystphoma havanensis, and Spermothamnion investiens. The Zonaria variegata and Polysiphoma havanensis were found only once and may have been brought here by the Gulf Stream, but the other three species were not uncommon and may confidently be believed to have come from the coral reef offshore. Codium tomentosum was collected in December and April but not in the intervening months, while Dictyota was not found there until after its occurrence in the harbor in June. Data concerning the condi- tions and alge occurring on the reef in winter would be of considerable interest, since it seems very probable that several species persist there throughout the year. 110307°—21——26 396 BULLETIN OF THE BUREAU OF FISHERIES. A comparison of the seasonal distribution of the Beaufort species which are found at Woods Hole and at Naples is given below, the numbers referring to the number of species and varieties common to Beaufort found in the other localities at the respective seasons: Beaufort summer flora. Beaufort spring flora. Beaufort perennial flora. aie s P s Pp. Ss P a um- eren- . . uim- eren- . uum- eren- Winter ier Saul Winter.| Spring. ier: SENT. Winter. 7B Seay WWO0ds Halen earn: pccdaweciys ual erage oie 13 3 I I 16 Af) 2-eatasy 7 4 Waplesror Gorn see ners ewe 9 3 2 In this case many of the species recorded above for winter at Naples are found there from autumn through spring, and some of the species recorded for summer at Woods Hole are found there in spring and summer. In general, though, it will be seen that the relationships of the Beaufort flora are greater with that occurring at Woods Hole in summer and at Naples in winter. The relations between the occurrence of any single species and the temperature are, however, frequently different in different localities. This is shown below where there is given the seasonal distribution of five species at Beaufort, Naples, and Woods Hole, with the range of the average temperature, in degrees centigrade, recorded in each locality during the time of occurrence of each species: Beaufort. Naples. Woods Hole. (Chamnia parvnlacc castes en: April to December, 17.5 to | Autumn to spring, 25 to 8 i 27.9 to 11.1°. to 19°. 4 Chondria dasyphylla................ April to February, 17.5 to | Autumn to spring, 25 to 8 é Oat 27.9 to 9.6°. to 19°. Chondria tenuissima................ April to June, 17.5 to 25.2°... Summer to autumn, 20 to 27 | }July to October, 20.43 to 20.97 to 18°. to 15.26°. Dasya'pedicellata ly. 4.5 Ae! ay 2 Ee? 2 4 AES) 2 Oe ia Pro) Wt) IPD ves > a 9 ) ~ SSH SE JSS DS Be S > ys eo) ) D> 3 D>) D>») WYD Bb YM) wP 2) De , PPD >) D> Sumy) D »») Eap?D VEO RR BQ Re , 22 P, > PEED») > 2 IP? Bp) WI VY? BYU» 2 2S YD Roe y >)» I>» >») > ) yy» DIF) 3 sa PD) DP» > >)) »p Sol b> oy» Nip» > Bs a2. ee Hs, r = t. ~ BP J a a | »y >) J ; 3 cee" ? 1 B DD PP} WIR) > WW SD YWW> vod Yrs) PD» D> Tp ; > »)»D >» PD) y >» D ) a) D> yo))))» »» Y) a pry » > Wey Dy). ) ))) » > as > Y) Dd Dy Mm Dep yPWPYvDF Dy 2S BP > wee 2 1. 3D Pp) YOY)» DD D: ) > DY >) >)? >> DIM) 3 )) DDD DPDm. wits» St Pi) » ))) ay am << eK SD >> pip 1) Dow) D> > D) D >) > > ay) . + PPD II tmp» Pee) w Pye) > py PHD sy - . ) Diop » »> me) me WD) uo» ». yyy dad x > yey) >? PHB) PIIDY? DB 2D WD) PW)» » YRVDZW)2D DH Wy » D> “2D WY) Pa y») » )»DWOH»_ WH» Wry Dp wy 29» De By >) ) > BY D >> > Jy PY? 2M oP PD) OFF Dy 2 a2 DD »)? BPd)>. 7) Le BP eD > ye») > yD») OP FJ 2 Ds 9» ) SP DIY DPD? DD DDYAD VP Dy 7 —i >» By) dD) >) DPy> Dd YAY?» 4>))» ae yy 28 £2 —-¥ ‘Sp ys Ie ¥ 4 re) pp > ey yy py Vn ) yy fhe ») y PM Soy? iy DPE) DW ew) | ep 2) 2), eID BD _ Da) D2 >) > Dy DID) PIP LYE» YD IY WIP” AD DIM Y» >» yape? LO PUP D> DP? 2D DS yw» D. we)? WD) So) ye ed) > D I>). » wD) ID PO, DIP) VP DD y). BIBS) SS Piyy wie Be » >. D> »d, P)) BBS > DIS). DvD DD)? > aD ) ee I SD) Py ty PUPP) 2) FP 44 ee Bs >»D> DP WIP I) ® DPyawy> D> BVI 4. 2 >) DPD PI, wv? D> >> > 92D _ BD Wy» Wry») 3 IOP) D2 DD ) PP eB» » JED >» yy WIID> I Iye yz») DY 2D >» yD» © Epw>e DS ® Ey ® YE 1) DMD | PP)? Jy 5 Do») >» IWLYDy » YW) 2 » 2) ae > DP»: D> DED) D Ip Xi Dee 2. Pow BD Die. 2) > Daw? > 2 BMI?) ‘ay WIND L> WPM PS Bp % >») » LP YoY) ae > ee » a RP 2) BW) SD OID > a) TD)» PVD» i —»)) » IDE» > > »v> Miblb \» wx =? D> WYP» Py > Sh, AE)? D >>> EDD )P WED pip DB) ware? 2 > By) I Dy) YD) BD a “) ) Io bs > DY BM»y

» > Pw DP) YP DP DP wy Dp > DP Im D 2 2D PPS 25s fIs bts > ) DB >W) to Be Pnm@) whimsy) bs ~ wv Ly s »: BS Dy Hy a 4 >. “AP ~ Pa) —+: S 2 ~ < > y) eR) i =" or ye ) a) yam > * ) > x DJ 4-232 ve) ) > Le) “E> Yo ¥ at ir 2») >> b> SD ey ID 33 222 D> Vw» -d » ) Pp - ») YB) Jy 1D) > >) ie <6 + 2 2» ay D> 1M ee) PPS 2537-7 > 0S Je 4 PD) DIDDY? »» JOD)» Di p y py} 5 Db») »d a> >> _ J Dey DP), 2 aye 2) » dy >) (_2@ Sis 33 Lae Ye oe» 2. DDD) ) Fase od 3 >> DY PHP DP DW YD? YD Hy» wy DD »>- >| 2S) 23» »-P De kh» > BID WID Dv”) IW)»d >» —~¥ 2 > — “yoNees > mp ) 2») ay ns ES i) > ») » PRYVMDV Ree 1») » PS ? > eo ») TP VD dy, be) DID D PIOD)) >) DIWW D> Le? Pee) ID ey) 22 P: a LEP BW D> )wD)) 9 )) WIP » » D2») PID yD? YD yy > Lae Wey) \ a») 2» LY) Dy) » J? DLP) RPP)» I» :Y soi 2D) PD DD» MH »Y) WwW) fo Dp IP > |} D>») » Po] > yy) 1) DPD DYE yp Ly THD" 2» Wy. DS PPP DYDD DP WN? >) DPE DD DY? PWD IM Wy? >) Sey. i ey?) EDD) el) SP @ ee ee De 2 SYP DPWOV>» IOP iP DD pI) » De 1? >) mpEp » SDD Dw BIA a? DP ppm? Kgs Vv» >? LY Db Dy”) Sy Ss: bP DD PY -, ) D Dap?’? Dw EDP 4 a) a) wD) >> pe yyy. ) 2 po 2 2? Pwr) B) my» DD 4 d SPP WTP PTD _ Dov? ys > » ) JEBE)»> 5) YP) ») > Ww ye > D> pI > i>) a? D> D>) EPP »D PW 2 ¢ II)» yD ”) >» Ys Wp Pp *¥ De) noe) Bre BD)» ~ 2s ) 1 > DY wy 2 IPD WD PID >? BOM y - 3 ts: 2D) PP Yor) BW Liaw > YP DW» y > » > D)>D>°y » JD Did > »» PID») ‘ ee ee yD) »y» 33 > a. ae hs YP YD»? YD a > > < —~4 7) wae ae” ee » PD. IPD)? >) You>> > De) we 2) ee) > 2 DDD)? »>P WD PW D2 D'D DVD» LP We) Fy) DP >». v»D> 2 DY. Oy iy) DD» Diy ye), »F Di?) fr Yd 2. > », rh | » b> itm Taam Vai Vee Bis) SR 2 —. yy) SS ‘~ | WMA 3 9088 01017 7228