Magnetic Spherules in Deep-sea Deposits

Magnetic Spherules in Deep-sea Depsits HANS PETTERSSON and K URT FREDRIKSSON 1 FEW OF THE COMPONENTS ente ring int the sedime nts cvering the cean flr have attracted s muc h attentin as the magnetic spherules first described in 1876 by Sir J hn Murray. Struck by the resemblance f their surface structure t th at f irn meterites, Murray called them "csmic spheru les." The number he was able t extract frm ne quart f the depsit, by means f a small magnet, varied frm 20 t 30 in samples f Red Clay, whereas in the same quantity f Glbigerina Oze nly ne r tw were fund (M urray, 1876). Tgether with A. F. Renard, Murray later published (1897) a mre detailed descriptin f the spherules in vlume 4 f the "Challenger" Reprts, Deep Sea Depsits. The higher num bers fund in the Red Clay are ascribed t its much lwer rate f sedimentatin. Hping that the methd f cunting the magneti c spherules present in different kinds f depsits may affrd a sluti n.t the difficult prblem f determining the rate f deep-sea sediment atin, ne f us decided t include in the equipment f the Swedish Deep -sea Expeditin (1947-48) special crers f wider diameter th an thse therwise used frm the " Albatrss," viz., with an internal diameter f 90 mm. instead f the usual 46 mm. Fr varius techn ical reasns this widebre crer was nt used until we reached the western Pacific Ocean (Petterssn, 1956). Owing t the higher resistance ffered by th e 1 Oceangrafiska In stitutet, Gt ebrg, Sweden. M anus cript received Nvember 20, 1956. Th e fllwing investigatin has been mad e pssible thrugh generus gra nts frm His M ajesty K ing Gu staf Adlf VI Fund, fr which supprt we beg t ffer ur mst respect ful thanks. Als, we have twice received grants in supp rt f u r wrk frm Karin ch Herbert ] acbssns Fn d, fr which we beg t ffer ur crdial thanks. 71 sediment t a thi ck cring tube, and wing t the necessity f aviding t heavy a strain n the steel cable used when crin g, the length f the thick crer had t be limited t nly 6 metres, as cmpared t the 15 t 20 metres f the narrwer cring tubes. The length f the thick cres raised was, therefre, in general nly 5 t 5Y2 metres. On the ther hand, frm a sectin taken frm a thick cre abut fur times mre material was btained than frm a narrw cre f the same length. In all, a dzen f such thick cres were raised frm the cruise with the " Albatrss" thrugh three ceans. Owing t mre pressing wrk n ther material cllected durin g the cruise, the wrking up f the thick cres fr magnetic spherules had t be pstpned fr several years after ur return frm the expeditin. Hwever, in the mean while a yun g technici an,. T. Laevastu, then in the empl y f the Ocean graphic Institute in Gtebrg, was charged with carrying ut preliminary extractin experiments using parts f a narrwer cre f Red Clay raised frm the central Pacific Ocean. Fr this wrk an electrmagnetic extractr f high effi ciency had been btained frm th e well-knwn New Yrk firm f Frantz. Prtins f the sediment suspended in water were passed th rugh this extractr. Already in the preliminary experiments with this instrument its great superirity ver the primitive me th d fr extractin used by Murray was apparent, the number f sph erules extracted frm 1 kg. f Red Clay varying between a few hundred and a cuple f thusands. Attempts t estimate the effi ciency f the met hd were als carried ut by adding t sediment already extracted a cunted number f artificialmagnetic spherul es, made frm irn wire with the xygen flame, having slightly larger dimensi ns thugh than the

72 natural magnetic spherules fund in the sediment. The yield was fund t exceed 90 per cent. The results frm the preliminary investigatin, including a graph shwing the sizedistributin f the natural spherules, has been published by Laevastu tgether with Prf. O. Mellis f Stckhlm (1955), wh had given him valuable help in the micrscpical study f the spherules, Our first bjective when starting ur wrk in Nvember, 1954, was t perfect the methd f extractin, invlving als the preparatin f the sediment samples btained frm the cres. The details f this preparatry wrk, carried ut in the Oceangraphic Institute with the excellent help fmrs. Karin Rmlin, technical assistant, will be given in a future publicatin by ne f us. It will suffice t mentin here that we fund it necessary t pass an aqeus suspensin f a quantity f the sediment, weighing frm 200 t 700 grams (with the narrw crer the sample weight varied between 20 and 300 gr.), at least three times thrugh the extractr. Frm the magnetic particles thus btained, which cmprised als a large prprtin f nnspherical particles f terrestrial rigin (largely magnetite ), thse with strng magnetic prperties were separated ut by means f a small electrmagnet and then passed thrugh sieves f different mesh, separating ut the three size-classes: carser than 60 u, 60 fj t 30 u, and less than 30 fj in diameter (see Figs. 1 and 2). The prducts frm this fractinatin were then munted between glass discs in a manner suitable fr cunting under the micrscpe. This latter peratin was made by ne f us in Stckhlm, mre recently assisted by fil. kand M. Nilssn. The peratins required fr a detailed examinatin f the spherules, like plishing, X-ray examinatin, etc., were carried ut in the Mineralgical Institutin f Stckhlms H6gskla (Fredrikssn, 1956). T its directr, Prfessr S. Gavelin, we are much bliged fr his supprt and advice. We are als in- PACIFIC SCIENCE, Vl. XII, January, 1958 debted t Prfessr F. Hecht fthe II Chernisches Institut der Universitat Wien, fr kindly carrying ut fr us micr-analyses fr nickel and cbalt n sme f the spherules submitted t him. His results prved nickel t be definitely present in a percentage varying between 6 per cent and 15 per cent f the irn cntent. Recently Dr. A. Smales at Harwell kindly investigated samples f deep-sea spherules fr nickel and cbalt by means f neutrn-activatin in the pile. In rder t get cmparable results frm ne sample t anther, we fund it advisable t cncentrate the cunts n black spherules f a diameter exceeding 30 p: The spherules smaller than 30 fj are difficult t cunt and easy t verlk. Their cntributin t the ttal weight f the spherules appears insignificant. Spherules f the greatest size, i.e., frm 60 fj t 250 fj were als cunted. Their cntributin t the ttal number f spherules is nt great, but, wing t their large size, their cntributin t the ttal weight is cnsiderable. The fllwing table gives a survey f the cres frm which we have extracted and cunted spherules. COMMENTS CORE 71. This cre is f special interest, as it was raised frm the vicinity fchallenger Statin 274 (S 07 0 25' W 151 0 15') where the depth was 2,750 fathms r 5,030 m. The sediment in the surface has been characterized by Murray and Renard (1897) as Radilarian Oze with 3.89 per cent CaC0 3 The cre raised frm the "Albatrss" has a ttal length f nearly 10 m., frm which sectins f 26 t 76 em. in length were taken. The number f spherules per kg. f sediment, free frm salts and lime, varied frm a maximum f 1,400 near the surface t a minimum f slightly mre than 100 per kg. A secnd maximum f 600 spherules per kg. appears at the 10 metre level. Cntrl cunts made n halves f the sectin, cut lengthwise, gave fairly large variatins in the prprtin

Magnetic Spherules- PETTERSSON AND FREDERIKSSON TABLE 1 73 LEN GT H OF COR E NO. D1AM. OF CORE IN LATI TU DE LONGIT UD E DEPTH NU MBER OF CORE, MM. MET ERS IN METERS SAMPLES 71 46 10.0 S 7 38' W 152 53' 4990 16 90 90 5.2 S 3 21' E 174 12' 4830 49 90 B 26 0.3 S 3 21' E 174 12' 4830 1 92 90 5.2 S 1 20' E 167 23' 3960 25 133 90 3.3 S 11 33' E 91 26' 5200 18 187 46 9.5 N 33 59' E 31 02' 2500 16 17 46 N 43 28' E 7 22' 2030 1* 18 46 N 41 29' E 5 51' 2680 1* 87 46 N 2 23' W 173 50' 5560 1* 89 46 S 2 48' W 178 57' 5480 1* Sep arate sample frm the surface. f 1:2. The average number f spherules per kg. fr the whle cre is abut 300 (see the curve in Fig. 3). Accrding t radium measurements, made by Kr ll (195 5), in the sediment where cres 71 and 72 were taken, the rate f sedimentatin is taken t be between 1 and 2 mm. in 1,000 years. CORE 72. T his cre was raised frm the immediate vicinity f Cre 71, where the sediment had the same character. Its uppermst 3 metres, ut f a ttal length f 14 metres, have been examined fr spherules by Laevastu and Mellis (4). In Figure 4 the results frm the ir cunts are reprduced frm their riginal paper. In Figure 3 the number f sph erules is set ut fr the same levels in which Cre 71 has been investigated by us. The figures, hwever, have been reduced t 50 per cent t be in cnfrmity with ur cunts, since Laevastu and Mellis in clude d spherules frm diameters 10 fj. upwar ds in their cunts, whereas ur cunts in Cre 71 are limited t diameters ffrm 30 fj. upwards. Accrding t Laevastu and Mellis the spherules fless than 30 fj. in diameter mad e up less than ne half f the ttal number. CORE 90. This is a thick cre, 90 mm. in diameter. Its ttal length is, therefre, very mderate, r nly 5 metres. In its up per part the cntent flime is lw, less than 1 per cent. Belw the 190 cm. level the cntent f carbnates increases abruptly t abut 30 per cent, rising still further d wn t mre th an 80 per cent. The number f spherules varies greatly. Frm abut 3,300 per kg. near the surface, a maximum f 5,000 is reached between 80 and 90 cm. Belw the n cm. level the number f spherules per kg. f salt- and carbnate-free sediment is less than ' 1,000 (min imum 140 ), with the exceptin f a secn dary maximum f 1,300 in a depth fabut 280 cm. Tw parallel series f samples frm this cre were extracted. The results btained frm the first five samples in the first series (see full drawn curve in Fig. 5) were much lwer than thse frm the secnd series. Th is is pr bably due t the fact that the technique f extractin had nt been fully develped at th at time, s th at a certain number f the spherules may h ave becme crushed in the extractr and thus escaped the cunting. In certain samples it prved very difficult t disting uish between different types f spherules, f which the black shiny nes were primarily cunted. This difficulty may explain the str ng variatins fund by cntrl cunts frm identical levels, especially as regards the sample between 63 and 80 cm. In th e diagram in Figure 5 these variatins are indicated by hriznta l lines, uniting the highest and the lwest values fund at the level in questin. CORE 90 B. Th is was a narrw "pilt cre" taken frm the same vicinity as Cre 90. The amunts f spherules in th e lime-free p rtins in the up permst 4-12 cm. and 0-6.5 ern. f Cres 90 Band 90 respectively were fund t

74 PACIFIC SCIENCE, Vl. XII, January, 1958 FIG. 1. Spherules frm Cre 90, abut 30 t 60 micrns in diameter. be practically the same. The carbnate cntent was, hwever, much higher in the uppermst part f Cre 90 B, viz., 56 per cent as cmpared with the crrespnding 0-6.5 em. f Cre 90. CORE 92. This cre, raised frm a depth f nearly 4,000 metres clse t the Equatr, is als ne f large diameter (90 mm.) with a length f nly 5 metres. The sediment is a calcareus ze, its cntent fcac0 3 varying between 71 and 82 per cent. The uppermst 64 em. f the cre were missing, having been lst in transprt. The rate f sedimentatin can nly be tentatively estimated at 50 mm. in 1,000 years. The number f spherules extracted was very lw. Recalculated t 1 kg. f lime- and salt-free sediment, the numbers varied frm in a depth f 201-217 em. t slightly ver 200 between 64-81 em., 231--248 em. and 275-293 em. The average fr the whle cre was 80 spherules per kg. f sediment. (See diagram in Figure 6.) CORE 133. This cre, the nly thick cre (90 mm.) raised frm the central part f the Indian Ocean at a depth f 5,200 m., had a ttal length f 326 cm., frm which the uppermst 3 em. were missing. Dwn t a depth f 200 em. belw (he tp f the cre it had a mderately high cntent f CaC0 3, between 14 and 40 per cent, whereas in the lwer parts f the cre the lime cntent was nly between 5 and 10 per cent, especially near the lwer end where it varied ver 45 em. between 5 and 5.6 per cent. The number f csmic spherules extracted frm the uppermst 3-18 em. was high, viz., nearly 700 per kg., whereas in the lwer parts, at depths between 50 and

Magnetic Spherules- PETIERSSON AND FREDERIKSSON 75 FIG. 2. Plished sectins f black spherules. A, spherule frm Cre 71, abut ne meter belw the sediment surface. Magn etite, gray, with a metallic nucleus, white. D iameter f spherule, 30 micrns. B, spherule frm the Atlantic Ocean, abut 25 em. belw the sediment surface. Same type as A, with imperfect plish. Di ameter f sphe rule, ab ut 80 micr ns. C, spherule frm the same sample as A. Mag netite, gray, metallic nucleus, white. Diameter, 120 micr ns. 220 em., the num bers were lw, rising again t smewhat high er values belw 220 em. The average number fr the whle cre was 130 spherules per kg., whereas between 52 and 220 em. it was nly 30, cmpared t the average fr the uppermst 50 em. f 500 per kg. (Fig. 7). T ascribe these very remarkab levariatins in th e numbers f the spherules nly, r even mainly, t changes in the rate f sedimentatin seems unwarranted. Tha t even in the uppermst layers the rate f sedimentatin was higher than in the central Pacific Ocean is indicated by radium measurements in a neighburing shrt pilt cre, Number 133 B, which cntained n an average 11 units f the 12th decimal place f Ragr/ gr, cm pared t frm 40 t 50 units in Red Clay frm the central Pacific. Th us the rate f sedimentatin can be estimated fr the upper parts f Cre 133 at 10 mm. in 1,000 years. CORE 187. This cre, raised frm a depth f 2,500 metres in the eastern Mediterranean suthwest f Cyprus, was a narrw ne and had a ttal length f 9Y2 metres. Hence fairly lng sectins f 30 t 70 em. had t be used fr extracting the spherules. The lime cntent was mderately high, varying bet ween 20 and 40 per cent f CaCOa. The number f sph erules per kg. f lime- and salt-free sediment varied cnsid erably alng the leng th f the cre, frm 190 near the surface t an absl ute maximum f 1,130 in the sectin 165-235 em., declining frm there, bth upwards and dwnwards, t 5 and 10 per cent respectively f th e maximum value. At anther maximum, in sectin 438-506 em., 491 spherules were fund per kg. f sediment. In the very lwest parts f the cre, 860-950 em., nly 38 and 49 spherules respectively were fund. It mu st, hwever, be emphasized that where cntrl samples frm

76 PACIFIC SCIENCE, Vl. XII, January, 1958 Spherules per kg salt-and carbnate-free sediment. 200 ~O 600 ' 800 O~~~~~=il<iii 1 em ~ 200 c.!: s: is.,!!; 400 600 800 1000 CORE 72 1400 FIG. 3. Number f spherules per kg. f dry substance frm Cre 71. the same levels were taken, the values varied cnsiderably. inter se. The average fr the whle cre is 190 spherules per kg. Regarding the rate f sedimentatin in this cre vety little is knwn. O. Mellis (1954), frm a study f the vlcanic ash hrizns present in Ctes frm the eastern Mediterranean, has identified a layer f vlcanic ash in Cre 187, situated at a depth f40 em. belw the tp f the cre. This layer prbably wes its rigin t the catastrphic utbreak f the island vlcan Santrin, which ccurred sme time between 1800 and 1500 B.C., i.e., abut 3,700 years ag. Th is assumptin wuld make the rate f sedimentatin in the upper layers f the cre abut 10 em. in 1,000 years, leaving aside pssible disturbances in the sedimentatin due t slumping, etc. Regarding the tw maxima fund in Cre 187, it shuld be nted tha t the spherules they cntain were especially difficult t identify, resemblin g sme f the samples in Cre 90. The values given here must, therefre, be stated with due reserve and have nly been.indicared in the curve in Figure 8. Nevertheless there is little d ubt that the maxima are real, even thugh the numerical finds are smewhat uncertain. Attempts t find such maxima in ther cres frm the vicinit y are at present being undertaken. Cnsidering the abrupt decrease in the number f spherules n bth sides f the principal maximum, ne is tempted t assume that at the time when the maximum ccurred ne r several meteritic shwers fell upn the eastern Mediterranean. An alternative explanatin, tha t at the time when the maximum ccurred there can have ccurred an abnrmally lw rate f sedimentatin, appears less prbable. It wuld be mst interesting t investigate ther cres, preferably f larger diameter, taken in the vicinity f Cre 187. MAGNETIC SPHERULES FROM THE ATMOSPHERE Th e prblem f finding by direct bservatin the amunt fmagnetic particles settling SPHERULES/ kg. DRY SUBSTANCE 500 1000 1500 2000 50 100 ~ISO ~ ~200 a. ~ 250 300 FIG. 4. Number f spherules per kg. f salt- and carbn ate-free sediment frm Cres 71 and 72.

Mazneric Spherules- PETTERSSON AND FREDERIKSSON 77 Spherules per kg slt- nd cqrnte-free sea'tinenl /000 2000.1000 /;000 5000 Or------,------,--------.---.=---,.------r-- em 100 200.100 FIG. 5. Number f spherules per kg. f salt- and carbnate-free sediment frm Cre 90. n the Earth's surface has repeatedly been attacked by varius investigatrs, mainly in the United Srates f America. The methd adpted has generally been that f intrducing magnetic "cllectrs" int the water sputs carrying rainwater frm extensive rfs. The number f magnetic spherules accumulated in this manner has been cunted with results which ' are surprisingly large. Thus Warren J. Thmsen (1953) finds the weight f such magnetic spheres, frm bservatins made at Iwa City, t crrespnd t a ttal mass f 2,000,000 tns per year fr the whle Earth. Still higher values have been reprted by W. D. Crzier (1955) wrking in New Mexic, viz., 35,000 tns per diem during August 1955, r nearly 13 millin tns fr the whle year. These figures are several thusand times higher than thse btained frm ur cunts f the spherules in deep-sea depsits. Hwever, neither f the authrs mentined culd find any trace f nickel in the magnetic spheres analyzed. We have tried t cllect magnetic spherules by means f magnetic cllectrs intrduced int water sputs munted at Brne Statin in the G ullmar Fjrd, n the west cast f Sweden, and als frm the rf f the Oceangraphic Institute in Gtebrg. At the latter lcality we fund the number f magnetic spherules cllected t be very high, whereas at Brn the number was cnsiderably less but

78 PACIFIC SCIENCE, Vl. XII, January, 1958... em c 100 u.!: x: Ci. ~ 200 300 400 500 I pari _~:sing CORE 92 FIG. 6. Number f spherules per kg. f salr- and carbnate-free sediment frm Cre 92. still higher than the numbers fund in deepsea depsits. N nickel was fund t be present in these spherules. We infer, therefre, that mst f the spher ules thus cllected are f terrestrial rigin and are artifacts frm industrial plants and welding peratins, which can be carried ver large distances by wind. Hppe and Zimmerman (1954), cllecting magnetic spherules at J ena and ther lcalities in West Germ any, have arrived at the same cnclusin and cnsider it imp ssible t avid such cntaminatin in lcalities in r near indu strial dist ricts. Based n these results we have planned t cllect magnet ic particles frm the atms phere n cean islands situated far frm any industri al plants and as remte as pssible frm the shipping lanes acrss the sea. Such measurements, which we hpe t rganize n Pacific Ocean islands durin g the Gep hysical Year, may be expected t affrd evidence f true extraterrestrial spherules settling ver the cean surface and thus cmplement ur studies f the magnetic spherules in deep-sea depsi ts. Calling attentin t the interesting results 'frm the attempts at a magnetic drag ver the cean bttm frm the " Galathea" Expeditin, we cnsider the evidence in favur f the magnetic particles reprted by A. Bruun and his c-wrkers (1955) as bein g largely f csmic rigin t be 'very strn g. On the ther han d this methd f using a superficial cllectr can nly be expected t bring up extraterrestrial particles f recent rigin. Mrever, a certain risk f cntaminatin with spur ius spheru les frm cal- r il-driven ships passing near the curse f the " Galathea " cannt be altge ther excluded. N such risk's are invlved when extracting magnetic spherules frm undisturbed lng cres prvided ne gets belw the upperm st few centimeters f sediment. Th is methd alne seems t allw f definite cnclusins regardi ng the frequency f meteritic falls and their variatin s in the remte past. Quite recently the accruement f meteritic fragments and dust t the surface f the Earth and its ceans has attracted the interest f meterlgists and climatlgists. M ainly thrugh the wrk f Bwen in Australia (1953) and f his c-wrke rs in Spherules per kg sail -and carbnate-free sediment. 200 400 600 O..------.--r--,--,--,---,--;r em ~ 100 u.!: s: ~ 200 c 300 CORE 133 FIG. 7. N umber f spherules per kg. f salt- and carbnate-free sediment frm Cre 133.

Magnetic Spherules- PETTERSSON AND FREDERIKSSON 79.. Spherules par kg salt-and carbnate-free sediment. 200 400 600 Or----,,...--:r--,--,.---,---r- em.. 200 j '00 (~- 1000 CORE 187 e- -1100 e-500 FIG. 8. Number f spherules per kg. f salt- and carbnate-free sediment frm Cni 187. different parts f the wrld (Bracewell, 1954) strng evidence has been prduced fr the influence f csmic dust as cnden satin nuclei fr raindrps. Accrdi ng t these finds there ccur distinct maxima f precipitatin n certain dates like January 12, 22, and 31 t February 1. These peaks f rainfall ccur 29 t 30 days after the Earth has passed thrugh met er shwers. Accrding t Bwen the dust int which these meters are dispersed at great heights in the atm sphere takes abut 30 days t descend t the upper trpsphere, where a cndensatin f waterdrps is prd uced which ultimately reach the Earth as rain. Thrugh this cndensatin a depletin fwater vapur in the upper atmsphere is ccasined, decreasing the "greenhuse" effect n the radiatin balance. It has even been surmised that an abnrmal increase in the incidence f meters may give rise, by a trigger actin, t a deteriratin fclimate and pssibly release a glacial epch. N wnder that students f the atms phere and its radiati n balance are taking a grwing interest in meteric dust. THE ANNUAL ACCRUEMENT OF COSMIC SPHERULES In trying t wrk ut the ttal weigh t f th e csmic spherules settling n t the surface f ur planet th e numbers f spherules fund in the cres investiga ted can be used, cnverting the numbers f spherules per kg. int weigh ts. But already thi s cnversin frm numbers int weigh ts implies an element f uncertaint y. We have cunted all spher ules f a diameter exceedin g 30 JJ. and then cmputed the ttal number accruing t the Earth fr ne year. These values are set ut in clumn 7 f Table 3. In cnverting numb ers int weights we have assumed a specific weight f 5.0 and a diamete r f 40 JJ., assuming that all the spherules belng t the size-class 30-60 JJ.. When taking the ttal cu nted number instead f the number f the size-class 30-60 JJ., we add abut 10 per cent as a crrectin fr the weight f the spherules f a diameter less than 30 JJ.. In this way the values set ut in clumn 8 have been btained. Finally, clumn 9 shws values f the t tal accruemen t f black spherules f all sizes, the value being three times higher than ths e in clumn 8. The factr 3 we btain by cnsidering that the number f spherules between 30 and 60 JJ. is 8 times higher than the number fspherules exceeding 60 JJ. in diameter. (See Tab le 2.) The weight f the spherules f 30-60 JJ. is assumed t be equal t the weigh t f the fractin 60-125 JJ. and 125-250 JJ.. CORE NO. TA BLE 2 TO TAL NUMBERS OF SPHE RULES COU NTED 30-60 p. > 60 p: N UMB ERS 30-6 0 P. NU MBERS 60 P. 71 349 69 5 90 2,634 303 9 92 115 23 5 133 326 29 11 187 507 58 9 3,931 482 8

80 PACIFIC SCIENCE, Vl. XII, January, 1958 Anther still greater uncertainty ente rs int the calculatin wing t the imperfectly knwn rate f sedimentatin. Here ne has t use apprximate values invlving pssible errrs by the factr f 2 t 5, in sme cases even mre. The fllwing values wrked ut at the present stage f ur investigatin must be given, therefre, with due reservatin. Cmpared t earlier figures given by ther investigatrs ur estimates are very mderate but are necessarily subject t an revisin when mre extensive investigatins nw pending have been carried ut. In general ne gains the impressin frm these tables that the meter frequency has been cnsiderably abve the average in recent times, as is als indicated in mst f the frequency diagrams. Hwever, we musternphasize the great difficulties inherent in the methd f sampling. First, the surface layer may be missing, a layer which may quite well represent a sedimentatin time f tens f thusands f years. Further, wing t different rates f sedimentatin, a whle cre like N. 187 prbably represents a sedimentatin time f less than 100,000 years, which crrespnds nly t a cuple f em. in Cre 7l. Hence, it is desirable t investigate a series f cres frm different lcalities in rder t arrive at really dependable values fr the accruement f spherules and by this means t interrelate the results btained frm different cres. TABLE 3 Apart frm the results given in the diagrams and in Table 3 we als have values frm fur ther samples, tw frm the upper layers f tw western Mediterranean (Cres 17 and 18), and tw frm the western Pacific Ocean (Cres 87 and 89). Frm Cres 17 and 18 the number f spherules per kg. lime- and salt-free sediment is 260 and 150 respectively. Assuming that the frequency f spherules is inversely prprtinal t the rate f sedimentatin at a certain time we can, by cmparisn with Cre 187 frm the eastern Mediterranean, frm an estimate f the rate fsedimentatin in the western Mediterranean. We thus find a sedimentatin rate f abut 80 mm. in 1,000 years fr Cre 17 and a rate f 130 mm. in 1,000 years fr Cre 18, which are bth quite reasnable figures. Fr Cres 87 and 89 the numbers f spherules were 1,725 and 1,165 respectively, and thrugh a cmparisn with Cres 90 and 90 B we btain a sedimentatin rate f 6 and 8 mm. respectively in 1,000 years. These figures are three t fur times higher than thse fund by Krll (1955) frm radium measurements. Hwever, we d nt knw hw far the uppermst samples in different cres represent the same span f time. SUMMARY The present paper gives the results frm a study f the number f " csmic spherules" present in sediment cres btained frm great depths in different parts f the cean. ANNUAL ACCRUEMENT OF BLACK SPHE RU LES Rate f Number f Number Weight Ttal Cre Depth sedimentatin CaCO a Span f time diam. > 30 (> 30,..) in tns weight N. in em. mm /looo y per cent in years per kg. (10Ir,) «60,..) in tns 187 0-39 100 35 z- 3,700 190 4.7 790 2,400 90 0-6.5 3 1 0?- 20,000 3,3 50 1.6 265 800 90B 4-12 6 56 7,000-20,000 3,3 50 1.6 270 800 92 64-81 * 50 76 13,000-16, 000 220 1.1 190 600 133 3-18 10 15 3,000-18,000 700 1.0 160 500 71 0-26 1.5 1 0?-200,000 1,400 0.31 60 175 72 0-300 - - -- - - - 125 t * Sectin 0-64 em. lst in transprt. t Accrding t Laevast u an d Mellis.

Magnetic Spherules- PETTERSSON AND FREDERIKSSON 81 The number f such spherules btained by means fa pwerful electrmagnetic extractr is tens f times greater than the numbers fund by Sir Jhn Murray and A. F. Renard by mre primitive means. Spheru les in cns iderab le numbers are fund als in depths f several metres belw the sediment surface, i.e., in sediments already depsited in Terti ary Time. This definitely refutes the hypthesis tha t meteritic falls, frm which the spherules are assumed t be derived, h'ave been limited t a relatively recent past, the last 25,000 years. On the ther hand there are strng indica tin s that the frequency f such spherules depsi ted in recent times, say in the last few thusand years, has been higher than in a mre remte past. There are als indicatins f intermediate maxima which may crrespnd t a greater frequenc y f meteritic falls during certain times cvered by ur material. A system atic crrelatin f such maxima inter se has nt been made. Cm paring the freque ncy f spherules fund in recent times, i.e., near th e surface, in certain cres affrds means f estimating the apprximate rates f sedimentatin. Cnverting the number f spherules fund per kg. f sediment int weigh ts makes it pssible t estimate the ttal accruement f spherules t the wh le Earth, the results being a mderate figure f 2,400 metric tns annually (with a pss ible rise t 5,000 tns), i.e., nly a small fractin f figures btained by earlier investig atrs frm the num ber f magnetic spherules cllected frm the atmsphere. Means f btaining mre reliable figures fr such spherules by establishing cllectin statins n mid-cean islands are indicated. In cases where the num ber f sphe rules frm deep -sea cres have been analysed fr irn, nickel, and cb alt, the presence f nickel has given a definite prffthe csmic rigin f the deep-sea spherules. REFERENC ES BOWEN, E. G. 1953. A ustralian J ur. Physics 6(4): 490-497. BRACEWELL, R. N. 1954. The Observatry 73 (877) : 249-250. BRUUN, A. F., E. LANGER, and H. PAULY. 1955. Deep-Sea Research 2: 230. CROZIER, W. D. 1955. Paper presen ted befre the Meterritical Sciety, September 12, 1955. FREDRIKSSON, K. 1956. Nature 176: 32. HOPPE,1., and H. ZIMMERMAN. 1954. N aturw 41: 93. K ROLL, V. ST. 1955. Reprts Swed. Deep-Sea Exped., Vl. 10, fase. 1. LAEVASTU, T., and O. M ELLIS. 1955. Am er. Gepbys. Unin, Trans. 36: 385.. M ELLIS, O. 1954. Deep-Sea Research 2: 89-92. M URRAY,]. 1876. Ryal Sc. Edinburgh, Prc. IX : 258. M URRAY,]., and A. F. R ENARD. 1891. Reprts "Challenger" Exped. 4: 327. PETTERSSON, H. 1956. Reprts Swed. Deep-Sea Exped., Vl. 1. THOMSEN, W. ]. 1953. Sky and Telescpe, p. 147, April 1953.