Chemical Analysis of 17th-century Red Glass Trade Beads from Northeastern North America and Amsterdam

. . \ n l t t t t ( , u t ( t r \ 4 3 . - 1 t l t X ) l ) 5 0 . 1 - 5 1 5 . P l r n t e d i r r C r c l r t B l i t a i n

CHEMICAL ANALYSIS OF ITTH.CENTURY RED

G L A S S T R A D E B E A D S F R O M N O R T H E A S T E R NN O R T H A M E R I C A A N D A M S T E R D A M *

M . L . S E M P O W S K I a n d A . W . N O H E

Roc:he.ster Museunt & Science Center. Rochester. New York 14603, USA

R . G . V . H A N C O C K

SLOWPOKE-2 Facilin, Department of Chemistn' und Chemic'ul Engineering,

Royal Militari ' Collrgr, Kingston, Ontario, Conuclu. K7K 781

J . - F . M O R E A U

Ddpurtement des sciences humaines, Uni 'ersitd du Quibec'd Clic:outimi, Chic'outint i , Qudber' , Canado, G7H 28l

F . K W O K a n d S . A U F R E I T E R

SLOWPOKE Reoctor Fcrcilin* cmd Deportment rlf Chenical Engineering antl Applietl Chenri,strv.

University of Toronto, Toronto, Ontario, Canada, M5S 3E5

K . K A R K L I N S

Park.s Canode, Ontorio Sen,ice Centre, Ottau'a, Ontario, Canudcr, KIA 0M5 q

J . B A A R T

Archaeologie, Stedeli.jk Beheer, Amste rclcun, The l"letherlantls

C . G A R R A D

Pettrn Reseorch Institute. Torortto, Ontario, Canada

and I . KENYON lDeceased]

Ontcrrio Heritupe Fountlulion. Toronto. Ontario, Canado. M5C 1.13

Seventeenth-centur\'opoque red (redvt,ood) gla.ss tade beads of dif.ferent shopes and .si7es

v'ere nrcttle of mi.recl alkali (mainlv soda)-lime g/as.se.s and were <'oloured with Cu,

pre.surnabl\'os cLtprous o.ride or as Jlnely di,sper,sed elementol Cu. During the earlv ITth

centun'. beacls of all shapes w'ere opaciJied vt' ith Sn; cored beocls, u,ith unc'olourecl cores and

hence low'er Cu let'els, also tended to haye .slightlt lrnter Sn conterils than unc'orecl beacls. By

the mid- lTth centurt,, cored tubtilar bead.s w*ere being opocifed u'ith a c'ontbination of Sn and

Sb, a technological r:hange ,sinti lar to that obsen,ed in white glass rrade bead.s, v'hile uncorecl

redv'ood beads (rppeur not to have been opaciJiecl vvith either Srt or Sb. Beud c'hemistries are

.su.fficiently, di.ffe rent to ollou' thent to be sortecl into subgroups, whic'h may then he trackecl in

v'urirnts urc'haeo\r.tgical .sites und regiorts.

KEYWORDS: RED (REDWOOD) GLASS BEADS. ITTH CENTLIRY. AMSTERDAM.

INSTRUMENTAL NEUTRON ACTIVATION ANALYSIS. SENECA. PETUN" ALGONQUIN.

NEW YORK. ONTARTO. QUEBEC

' r ' R c c c i r r ' t l l 8 N o r c n r b r : r 2 0 0 0 : a c c c p t c - r l - 5 F e b r L r u n l ( X ) 1 .

i ( - . i ,n ivcrsrn o l Oxl i r r t l . l00 l

504 M. L. Sempowski et al.

I N T R O D U C T I O N

The production and trading of glass beads by the Dutch and the French (and possibly theEnglish) during the 17th century resulted in vast accumulations, including opaque red beads, inarchaeological burial and habitation sites in northeastern North America. Kidd and Kidd's(1970) bead classification scheme, expanded by Karklins (1985a), which allows the typologicalclassification of these beads by their physical attributes, indicates a number of different typesand varieties of red (redwood) beads in these assemblages.

During the period under study, native groups in northeastem North America were acquiringEuropean-made glass beads, along with other types of manufactured goods, in trade or exchangeeither with European merchants, or with other native groups. The Dutch, with a semi-permanenttrading establishment along the Hudson River during the 17th century, were a major supplier tothe Seneca and other Five Nations Iroquois groups in New York, and it seems likely that someDutch-traded (and presumably made) beads may have been reaching Iroquoian and Algonquiangroups in Ontario and Quebec during this period. There was a brief period, however, in the mid-1660s when the English took over control of Fort Orange and the commerce that took placethere.

Along, and north of, the Saint Lawrence River, the French brought beads with them. Althoughthere is no definite evidence as to the origin(s) of these beads in Europe, some evidence points tothe existence of bead crafting in France at least from the onset of the 17th century. Moreover, itmay well be that the dynamics of exchange along the Saint Lawrence River itself, given theproximity of Iroquoian groups, was of a different nature from that prevailing in the Algonquianhinterland north of the Saint Lawrence where, from Lake Saint Jean, westetn rpgions as faras Huronia and Petunia could be easily reached through a 'nofthern route' (see Trigger 1981,Moreau 1994,32).

To date, there has been no systematic study of the elemental composition of 17th-century redbeads or their distributions, although there have been extensive, chemical analyses of blue glassfrom France (Gratuze et al. 1992, 1995, 1996; Soulier et al. 1996), blue glass beads fromnortheastern North America (Hancock et al. 1994; Kenyon et al. 1995; Hancock et al.'1996,Moreau et al.1991; Hancock et a|.2000), and white glass trade beads from northeastern NorthAmerica (Hancock et ol. 1991, Hancock et al. 1999' Sempowskr et al. 2000). This studyattempts to address that deficit by comparing the chemistries of red glass beads and fragmentsfrom a glass beadmaking house in Amsterdam with the chemistries of beads from Petunia insouth-central Ontario, from Seneca territory in westem New York State, and from Ashuapmush-uan in Quebec (Fig. i ).

E X P E R I M E N T A L P R O C E D U R E

Two hundred and twenty-one red glass beads of different shapes and sizes were collected forcomparative pulposes from a glass beadmaking house in Amsterdam (A) (65), from Petun (P)

sites in south-central Ontarto (31), from Seneca sites in New York State (N) (102), and from theAlgonquian site of Ashuapmushuan in Quebec (Q) (17). These glass beads were analysed non-destructively using instrumental neutron activation analysis at the SLOWPOKE Reactor Facilityof the University of Toronto. Like Cu blue glass beads, these beads needed to be irradiatedenough to obtain relatively sensitive measurements of As and Sb (Hanco ck et al. 1994, Kenyone t a | . 1 9 9 5 ) .

Beads of mass 5-l0mg were first cleaned ultrasonically, as required. They were storedindividually in 1.2 ml polyethylene viais, and were irradiated serially for 5 min at a neutron flux

Chemical analysis of red glass trade beads 505

of 1.0 x 10'' neutrons cm -s-' . Five to seven minutes after irradiation, the induced radioactivitywas counted for 5 min using a hyper-pure germanium detector-based gamma-ray spectrometer.This produced analytical concentration data for cobalt (Co), tin (Sn), copper (Cu), sodium (Na),

aluminium (Al), manganese (Mn), chlorine (Cl) and calcium (Ca). The samples were recountedfor 5-33 min the next day to measure the concentrations of the longer-lived radioisotopes ofsodium (Na), arsenic (As), antimony (Sb) and potassium (K). The sodium measurements wereused to link both counts. Elemental concentrations were calculated using the comparatormethod. Beads of larger masses were iradiated at suitably lower neutron fluxes to makeenough radioactivity for reasonable chemical analyses.

THE ARCHAEOLOGICAL CONTEXT OF THE BEAD SAMPLES

Amsterdam gloss beadmaking house

A total of 65 tubular bead fragments from an early lTth-century glass beadmaking house(Asd/Kgl0) in Amsterdam cbnstitute the Dutch bead collection. The samples consist of brokentubes and fragments representing waste discarded in the manufacturing process. Rescueexcavations at two contiguous sites (Asd/Kg9 and Asd/KglO) on the Keizersgracht canalbetween Wolvenstraat and Hartenstraat in downtown Amsterdam uncovered extensive glass-

and beadmaking waster deposits (Karklins 1985b). The larger of the two sites (Asd/Kg10) wasuncovered in 1981 by the municipal Department of Public Works while installing a sewer linealong the east side of the canal. Glassmaking debris was found to cover a region at least 70 mlong. Recovered by carefully washing and sorting the fill, the beads and beadmaking refuse wasfoundtobeconcentrated in an area about2m across and at a depth of 4m below street level.Over 50 000 specimens were recovered, including finished beads, malformed rejects and tubefragments, all of drawn manuf-acture. The deposit also contained chunks of varicolored glass,

numerous coloured glass rods, crucible remnants and a variety of drinking-glass fragments.all indicating that this was material from a local glasshouse. The archaeological context, aswell as the associated ceramics and slassware. reveals that the glasshouse operated between1601 and 1610.

Ontorio sites

The Ontario glass bead samples (37) derive from six Petun sites located in central Ontario. southof Georgian Bay and east of the Niagara escarpment (Fig. l). These Iroquoian village sites wereoccupied during the mid-17th century (c. 1630-50), and were presumably abandoned no laterthan 1650 (Thwaites 1896-1901,56, 115), when the Petun finally dispersed following severalyears of hostilities with the New York Iroquois.

New York sites

The New York bead samples (102) were recovered from 1l Seneca Iroquois sites that wereoccupied from the early to the late 17th century in western New York State (see Fig. l). Duringthis period, there were two distinct groups of Seneca, each occupying a separate village in closeproximity to the other, pius one or more small associated villages. Like other northern Iroquoianpeople, the Seneca abandoned their villages and moved approximately every 15-20 years,usually to a new site only a few miles away. The result is two parallel sequences of villagesites-an eastern and a western series. The Seneca were the first Iroquoian group whose

ctT__*_---j--( , ' . . H u d s o n B a y

506 M. L. Sempowski et aL

Figure I A malt ol northeastem Nrtrtlt America, sltov'ittg tlte regiotts of tlte Senec'u, Petutr ontl Algonrluian peoplles.

sequential movements during the period immediately following European contact were tracedarchaeologically (Houghton 1912, 1922; Wray and Schoff 1953; Wray 1913); these earlyformulations have undergone only slight modifications since that time (Wray et al. 1981, l99I;Sempowski and Saunders 2001). Furthermore, the unusual continuity of the sequence of Senecasites for the early historic period has made it a benchmark for comparative studies ofarchaeological assemblages from other areas.

Estimates of occupation dates for the Seneca sites in the two series are somewhat moreproblematic, but a few key historic events and changes in assemblages of European manufac-tured goods have permitted approximations of beginning and ending dates for each pair ofcontemporary sites. While further refinements are inevitable, the present sequence andchronology of Seneca sites appears to be close to the mark, with errors likely to be within5-10 year margins.

Quebec site

Finally, a total of ll glass beads from the Algonquian site of Ashuapmushuan in Quebec werestudied (see Fig. 1). They were found in a pre-1700 stratigraphic context on the site, although theexact dating of this earlier occupation level is not secure (Moreau et al.1991).

76(,

Chemical analysis o.f red glass trade beads 507

R E S U L T S A N D D I S C U S S I O N

The beads are all opaque red (redwood) (Kidd and Kidd l9l0) and all are soda-rich mixedalkali-lime-silica glasses (see Table 1). The Na contents range from 7.0 to l0.7Vo, the Kcontents varying inversely from 5.6 to 0.l%a. The Ca contents range from 2.2 to 8.57o, with thevast majority of beads having over 4.57o Ca. The beads were coloured red with Cu (with levelsranging from 0.4 to 2.8o/o), and were opacified primarily with Sn (0.4-3.47o), sometimes with Snand Sb (<0.87o), and less frequently with Sn and measurable low amounts (< 0.3Vo) of As. Basedon relative amounts, the Sb additions appear to be deliberate, while the As additions wereprobably accidental.

Given the well-dated site contexts of the Seneca beads included in this study, a number ofgeneralizations may be made about the chemistries of different types of redwood glass beads inuse during various periods in the lTth century. Four mutually exclusive, broad groupings areidentifiable on the basis of chemical profiles. Two of these groups relate primarily to beads fromsites dating to approximately the first half of the lTth century, and two groups relate to beadsfrom sites of the second half of the 17th century. The complete data for individual samples areavailable from the authors.'

Pre-1655 beads

Two broad chemical groups are distinguishable among beads dating to the first half of the 17thcentury. These include all but four of the redwood beads from Seneca sites dating beforeapproximately 1655; all of the beads from the pre- 1651 Petun sites in Ontario, and all of the beadfragments from the 1600-10 Amsterdam glass beadmaking site. The two groups db not,however, provide clear and unequivocal differentiation on the basis of the technique of beadmanufacture (cored and uncored), or on bead shape (round, circular and tubular), or ondecoration (striped or not striped). Nor does chronology offer an unassailable differentiatingfactor, since both groups appear to have been represented among the Amsterdam glass samples.Nevertheless. there are tendencies for certain varieties of beads to cluster within these twogroups.

Table I Sununcrr\ duta.for tlte .four broacl chemical groupittgs

Group l,I l1 samples

Group 2,14 santples

Group 3.23 santples

Grutup -1a.

.four samplesGroup 1,

36 sample.s

Al (7c)

Ca (c /c )

c t ( % )

Co (pprn)

Cu (%')

Mn (ppm)

K (7c)

Na (%)

Sn (7c)

As (ppm)

Sb (ppm)

0.90 -1_ 0.175 . 8 - F 1 . 1

0 .70 - f 0 .12<87 -r 790.93 -r 0.24

4900 -f 18003.1 -1_ 0.89.0 -1_ 0.9

1.71 -1_ 0.83<200 -r 220<12 +- 69

0.94 -r 0.124.8 -1_ 0.7

0 .68 - f 0 .10<65 -1_ 15

1.11 -r 0.223700 -'_ 900

3.2 -]_ 0.58.5 -]_ 0.9

2.65 "r 0.76<270 -r 200<110 - f 65

1 .07 - f 0 .166.0 -1_ 0.8

0.5,1 -f 0.05<88 -1. 490.68 - f 0.218600 -f 3800

3.0 - f 0.88.1 - ]_ 0.3

0.80 - f 0.34<210 - f 1003900 -f 2000

1.08 - f 0.056.3 -1_ 0.6

0.54 -f 0.04<93 + 141.31 - ' - 0.09

5700 -1_ 9004.2 -r 0.58 .0 - f 0 .1

2 . 3 6 - f 0 . 1 1<310 - f 2803000 -f 400

0 .80 - f 0 .116.6 -r 0.1

0.45 -f 0.08<53 -f 300 .90 - f 0 .13

4400 'f 21003.3 - f 0.88.8 - f 0.8

<0.14 - f 0.06<240 -r 90

<83 - f 7 l

'< . r I r " inro l ics that dctcct ion l in i t data w'ere inc luded in the calcu lat ion of the mean concentrat ion

508 M. L. Sempowski et al.

Group I beads (114 samples, summary data in column I of Table l) are characterized by lowCu (< l.5Vo), medium to high Sn (0.42-3.57o), and low to undetectable Sb (<700ppm).Group Iis the most eclectic in terms of bead shapes and manufacturing styles (see Table 2). Itenconrpasses the vast majorrty (897o) of the Seneca beads dating to the period from c. 1610to c. 1655 and l5Va of those in the Amsterdam sample. By contrast, only 4l%a of the Petun beadsare included in this group. The group includes all of the cored (round, circular, striped-tubularand barrel-shaped) redwood beads identified from sites of this period, as well as the majority(83To) of the uncored round and circular beads. A smaller percentage (40o/o) of uncored tubularbeads also falls into Group 1 : three from Seneca sites and 13 from Petun sites. For the most part,the cored beads tend to have lower Sn contents than the uncored beads. In terms of chronologicalrefinement, Group 1 beads occur primarily on Seneca sites dating throughout the entire first halfof the lTth century (i.e., from about 1610 to 1655). In addition. Group I also includes fourtubular (N71, N77, N79 and N88) and two round (N66 and N76) beads from later Seneca sites. Itis conceivable that they may have been holdovers from the earlier period.

Group 2 beads (44 samples, summary data in column 2 of Table 1) are distinguished fromGroup I beads by their higher Cu contents (>1 .57o), and on average, slightly higher amounts ofSn (0.69-3.5Vo). Like Group f . i ts beads are also characterized by low to undetectable Sbcontents (<300ppm).Group 2 includes only uncored beads (Table 3): primarily plain tubular(19). along with five (of six) square or twisted square tubular (all fiom Petun sites), and threerounded beads. Nearly 60Vo of all the Petun beads belong to this group, whereas only six beads(107a) of the pre-1655 Seneca sample belong here, all of them from sites postdating 1625.Finally, 16 beads, or nearly one-quarter of the Amsterdam sample, are included in th"is group.

Table2Asunmtan,tlfthedistrib,i,,,,"?!'uiiii,!,,',,i,,if,!,i,ii,i,,,,i::::;:,,,i,1:';;

Group De.sc'riptiort

21

21 cored round or c i rcular beads (pre-1655):

I cored striped tubular bead (pre-16-55): N502 cored striped barrel shaped beads (pre-1655): N44. 4-5

4 cored tubular beads (post-1655): N71. 77. 79. 882 uncored round beads (post-16-5-5): N66. 76

3 uncored tubular beads (pre-1655): N32. 10.4113 uncored tubular beads (Petun)

2 cored round beads (Petun)

17 uncored round or c i rcular beads: N05. 33. 48. 55. 21.

51 . - s2 , 54 . 14 , 16 . 11 , 29 .21 . 15 . 18 . 26 , 23

49 Amsterdam bead fiagments1 uncored round bead: P73

2l assorted uncored tubular beads (Petun)

3 uncored tubular beads (pre-1655): N36, 31, 46

3 uncored round or circular beads (pre-1655): N3 1, 34, 5316 Amsterdam bead fiagments23 cored tubular beads (post-1655)

4 cored tubular beads (1640-55); N62-6519 uncored round beads (New York, post-1655)

17 c i rcular beads (Quebec, 1625-1700)r9

Chemical analysis of red glass trade beads

Table 3 A summam o.f the Petun and Seneca beaels br- .shape

509

Group Urnd Ctube Utub CStrTub CStrBarrel

1 2

1 2

l 6

1 A

t 3 t ]2 4

- 4

A *

_ -1623 -59 3l

Pre- I 655Post- I 655Pre- I 655Post-l 665Pre- I 655Post- I 655Pre- I 655Post- I 6-55Petun * Seneca

1

Total 40

Crncl . cored round-shaped beads: LJrnd. uncored lound-shapccl bc-ads: Ctub. cored tubular-shapcd beads:

Utub. uncored tubular-shaped beads: CStrTub. cored. s t r ipcd t r - rbr- r lnr -shaped beads: CStrBarrc l . cored.

str-iped barre I -shaped bc.acls.

Post-1655 beads

Two additional, broad chemical groups encompass the redwood beads from Seneca sites whichdate later than 1655, and those from the Algonquian site of Ashuapmushuan in Quebec, alsothought to date to the latter half of the l Tth century. These groups appear to provide a clear andindisputable pattern of differentiation between the chemical make-up of cored tubular b6ads inuse during this period and that of uncored round and circular beads (see Table 3). Both groupsare characterized by relatively lower Sn levels than shown by either Group I or Group 2. In onecase. measurably larger amounts of Sb are combined with medium Sn contents in the productionof cored tubular beads. In the other, very low Sn contents combine with very low tounmeasurable amounts of Sb to produce uncored round and circular redwood beads. Both ofthese groups exhibit the low Cu levels typical of all of the groups identified here, except forGroup 2.

Group 3 (21 beads, summary data in columns 3 and 4 of Table 1) is distinguished clearly fromthe other broad groups by its relatively high Sb contents (1150-8500ppm). It is alsocharacterized by low Cu levels (<l.5Vo) and medium Sn contents (0.35-l.1Vo). This groupconsists only of beads from post-1655 Seneca sites in New York State and includes only coredtubular beads (see Table 3). Indeed, all but four of the cored tubular beads dating to Seneca sitesoccupied after 1655 (N71, N77, N79 and N88; see Group 1) fall within this chemical profi le. Asstated earlier. it seems likelv that these four beads mav have been holdovers from an earlierperiod.

Four cored tubular beads from the Power House site (Seneca, AD 1640-55; N62-N65, Group3,{, summary data in column 4 of Table I ) exhibit the higher Sb contents indicative of thisgroup, but with higher Sn readings typical of the earlier period. These four anomalous beads mayrepresent the vanguard of the Sb-rich beads.

Group 4 (20 beads, summary data in column 5 of Table 1) is distinguished primarily by itsvery low Sn levels (<0.357a), together with low Cu (< I.5Vo) and low Sb (< 300 ppm) contents.This final group includes only uncored round and circular beads from post-1655 Seneca sites inNew York State, and beads from the Algonquian site of Ashuapmushuan in Quebec. Indeed,only two (97o) of the uncored round and circular beads from Seneca sites dating after 1655 do not

5 1 0 M. L. Sempowski et al.

fall into this group. These two beads (N66 and N76) exhibit the higher Sn content typical ofGroup l , pre-1655 beads (1.1 and 3.42%, respect ively) .

Cu/Sn voriabilii- over time

Although red glass is known to be opacified by the presence of cuprous oxide or elementalcopper in glass (see, e.g., Freestone 1987), Sn was intentionally added to make redwood glasstrade beads in the early 17th century.

Table 4 shows the Cu and Sn variability over time for beads from the Petun, Seneca andAlgonquian sites, with the Amsterdam beads anchoring the early l Tth century. In the early l Tthcentury, both cored (relatively lower Cu and relatively lower Sn) and uncored (relatively higherCu and Sn) beads were produced. The Petun beads cover a short time period (c. 1630-50) andare much more consistent in their Cu and Sn contents, with Cu varying from 0.8 to 2.07o and Snfrom 1 .5 to 3.6Va (with one exception (P28) at <0.5Vo Sn). These variations are primarilyassociated with uncored tubular beads.

Table zl Copper and tin variations ot'er time

Period Si te Identific'ation l,{untber Vc Cu Vc Sn

r ' . 1600-1000 AmsterdamLow CuHigh Cu

Petun (Ontario)

c'. 1630-42 Hamilton-Lougheedc:. 1630-42 Connor-Rollingc.1639-49 McEwanc.1637-49 Kel ly-Campbel lr '. 1637-50 Plater-Martin

r . 1 6 3 7 - 5 0 Plater-Fleming

Senec'a (Nev' York State)

c . 1610-25 Dutch Ho l low

c. 1625-40

c . 1640-55

r ' . 1655-75

c . 1 6 1 5 - 8 7

P01P02P03, 04P05-09Pr 0-30

P 3 r 3 8

w NOl-0-5Nr 6-25E N06-15E N26-30w N 3 r - 3 5E N36-4.5w N46-5t)w N-5t -65E N66-75

w N76-90E Ngt 95

"\' N96- 10,5

N 1 0 6 - 1 1 5

Q 0 1 - 1 7

0.4-1.41 .5 -2 .2

0.92.0

1 . 5 - 2 . 11 .0 -2 .00.8-2.0

0.8- r .8

0.,5- 1.-50 .6 - l . t0.4- | .50 .9 - 1 .30 .1 - l . 10.6-1.50.1-2.00 .5 - r . 50 . 1 - 1 . 2

0. ,1- 1 .30.8-0.90.-5-0.70.1-1.2

0 .8 - 1 .0

0.4-3.3" 0 .6 -3 .1

3.4. A

J . +

3.0-3.51 .9 -3 .61.5-3..4

(P28 at <0.7)1.9-3.2

0.6-2.01 .2 -2 .10.5-2.41 . 4 - 3 . 11 .2 -3 .40.9-3.3I . ) t <t . _ - _ . - r

0 .6 -2 .9< 0 . 1 - 1 . 7

( 4 a t < 0 . 2 % )< 0.4-3.4

< 0 .30.7 -0.9

< 0 . 2

< 0 . 1

II24

2 1

518

Factorl,Hollou'CornishBosley Mi l lsWarrenMenzisPower HouseMarsh

DannRochester JunctionBoughton Hill

l 5t 0l 05-)

l 0-5

1.510

l-55

l 0l 0

Algonquian (Quebec)

r ' . 1625-1700 Ashuapmushuan

E. Eust Scncca: W. Wcst Se'ne-ca

t 1

Chemical analysis of red glass trade beads 5 1 1

The Seneca sites, with their longer time span and greater diversity of bead types. tell adifferent tale. It appears that prior to 1625 the beads, both cored and uncored, had low to mediumCu (0.4- L57a) and Sn (0.5-2.1Vo). Between 1625 and 1655, roughly the same time period asrepresented by the Petun beads, the Cu (0.5-2.57o) and Sn (0.6-3.47o) contents tended toincrease, showing an increased contribution of uncored beads. After 1655, the Cu reduced to itsearlier levels (0.4-1.3a/o), while the Sn, with the exception of f ive beads (N71, N76, N77. N79and N88), decreased. The Sn content becomes almost negligible in uncored round beads fromSeneca sites occupied after 1655. The elemental make-up of uncored round beads in Group 4,lacking either Sn or Sb as opacifying agents, seems to represent a major divergence from any ofthe other chemical profiles identified here. It may be possible to explain this drastic alteration inthe make-up of uncored round beads in terms of a change in the European source of the beadsacquired at this time.

In ,qo 1664, the English defeated the Dutch and took control of Fort Orange (Huey 1988). Onemight assume that the English were then able to enforce the Trade and Navigation Acts, whichrestricted trade in English colonies to English-made goods and English-owned ships, resulting inthe exclusion of Dutch goods on New York sites postdating that period (Eccles 1959, 9l).However, that result seems unlikely, because of Dutch strategies for circumventing therestrictions by renaming ships and acquiring English trading partners (Kupp I9l4). Furthermore,it has been suggested that the situation may have created opportunities for the acquisition ofFrench trade goods by New York merchants, because of an insufficient supply of Englishmanufactured goods such as glass beads prior to the 1680s (Paul Huey, personal communication2000). Thus it seems likely that, for a time-until the Dutch regained control of the fort atAlbany-Dutch-made glass beads may have been less available in the New York trade thal theyhad been in the preceding decades. The fact that beads on Seneca sites are so similar to those onthe Algonquian site at Ashuapmushuan in Quebec may indicate the French as the most probablesoufce.

Inter-site and inter-regiort connections

Table 5 lists some potentially similar chemistries of glass beads from different sites. Asexpected. in the l ight of the Wenro migration of 1639 to Huronia (Thwaites 1896-1901, 17,25-31 ). and to Petunia in 1649 (Thwaites I 896- 1901 , 39, 251), Dutch trade beads appeared atboth Seneca sites and Petun sites, in some cases the same glass bead chemistries being found inboth northeastern North American regions. In all cases of matching bead chemistries, the datesfor the North American sites are equal to or later than the operating period of the Dutch glassbeadmaking house, confirming the relative reliability of archaeological site-dating systems.Similar chemical connections are to be found between later Seneca and Algonquian sites. Theseinter-site bead-chemistry connections show that trade beads from Amsterdam reached differentcultural communities by combinations of relatively direct trading with Europeans together withtrading and migrations within aboriginal communities.

The Algonquian beads from Quebec are noteworthy. They were found at Ashuapmushuan,stratigraphically below a concentration of beads that was allocated a 1700-50 date (Moreau et al.1997). Typologically, they fit within the c. 1625-50 period. But, from their chemistry, theyappear to match two beads from the East Seneca Marsh site (c. 1655-75).

Even with the small numbers (36) of typologically matching low-Sn glass beads, it is temptingto speculate that the low-Sn glass beads travelled through Quebec (c. 1625-50) to thecommunities of the Seneca (post-1655), or possibly more directly, via French traders. Such asuggestion could be at least partly sustained by Sagard's relation of red bead trading (Sagard

5t2 M. L. Sempowski et al.

Table 5 Inter-site cttnnections

AI(Vc)

Ca Cl Cu(7") ("/") (7")

M n K N a S n(ppm) (Vr) (Vo) (ppm)

Amste rdam ( A ) - S ene c'o ( N ) c'onne cti ons

A30 1600-10 Amsterdam

A114 1600-10 Amsterdam

N2l c. 1610-2-5 Dutch Hol low

N5l c. 1640-55 Power House

A88 1600-10 AmsterdamA117 1600-10 Amsterdam

N,16 c. 1640-55 Menzis

A28 1600-10 Amsterdam

N01 c. 1610-25 Dutch Hol low

Nl6 c. 1610-25 Dutch Hol low

Amsterdam (A )- P etunia ( P ) connectionsA91 1600- l0 Amsterdam

Pl7 c. 1637-50 Plater-Martin

Seneca ( N)- Petunia ( P ) connectionsP04 c. 1631-49 McEwan

Pl4 c. 1637-50 Plater-MartinP31 c. 1637-50 Plater-FlemingN53 c. 1640-55 Power House

A109A 3 lN48 c.

01012 l

5 .05 .36 . 1

4.84.94.95 .76.66 . 1

0.851 .080.950.91l . l 01 .020.98

0 .81 4 .50.94 4.1

0.99 4.41.08 4.01 .10 4 .00.98 4.3

0.65 r .030.s4 1.250.64 l . r00.53 0.840.48 1.680.62 1.660.61 1.980.7 5 0 .850 .72 t . t ]0 .11 0.83

0 .53 2 .180.58 1.65

0 .82 2 .130.69 1.140.11 1.500.66 1.53

4210 4.33930 3.s4290 3.34240 4.42530 3.22170 2.82350 2.93220 3.53610 3 .63s20 2. '7

4590 2.84600 2.5

3630 3.33560 2.84160 2.83900 3.0

8.3 27 3007.5 210001.9 267008.0 26 8001.4 291008.0 269008.3 25 0009.1 23 s009.6 202009.9 24 800

7.2 21 5008 . 2 3 1 8 0 0

9.4 35 r008.8 31 6009.5 30 0009.t 29 400

Amsterdant (A)-Seneca (N)-Petunia ( P ) connections

P21N76At21P09N79N7 l

c . .

C .

L .

C .

( .

1600-10 Amsterdam 1.02 5.0 0.60 1.16 3210 3.3 7.9 310001600- 10 Amsterdam l. l5 5.5 0.60 1.03 3550 3.5 8.3 329001640-55 Menzis 0 .81 5.8 0 .51 1.42 3310 3.9 1 .6 23 0001637-50 Plater-Mart in 0.88 -5.6 0.59 1.43 3400 3.8 7.2 28 5001665-15 Dann 0.86 5.6 0.51 1.22 3630 3.5 7.4 342001600-10 Amsterdam l . l4 6 .0 0 .69 1.20 6690 3.8 10.0 184001639-49 Kelly-Campbell 1.04 5.2 0.62 1.03 6950 3.2 9.9 223001665-75 Dann 0.84 5.6 0 .55 0.69 6410 1.9 8 .8 l8 4001665-15 Marsh 0.16 6.7 0.66 0.69 6790 2.3 8.9 17 900

Seneca ( N )-Al gonquian ( Q) c'onttec'tiotts

Q05 r'. 1625-1100 Ashuapmushuan

Q10 r ' .1625-1100 AshuapmushuanN67 c'. 1665-15 Marsh

N70 c. 1665-15 Marsh

0 .81 6 .60 .82 6 .70 .84 6 .10 .79 1 .3

0.36 0.910.43 0.860.48 1.010.48 0.86

4140 3.64040 3.34010 3.54510 4 .2

9 .2 < 10009.2 < 8,509.6 < I -5008 . 5 < 1 4 0 0

1865,243 1866,129) . I f th is re la t ionmakes i tc lear tha t redg lassbeadsarepar to f the t rad ingbusiness of one specific Algonquian group of the Ottawa region, there is a wide array ofinterpretation concerning the nature of the business involved (see Kenyon 1986, 56), as well asits interpretation in terms of the orientation of the trade network (from Huronia eastward or fromSubarctic Quebec-such as Ashuapmushuan-westward; Moreau 1994. 33).

A second scenario would favour c. 1650-1700 as the period for the Ashuapmushuan beads.Given the likely disruption of Dutch trading items to the south of the Saint Lawrence River inthat time period, it is possible that the Senecas obtained their red, low-Sn beads by trade with theAlgonquian people localized north of the Saint Lawrence. It is also possible that they obtainedthem by trading with other Five Nations peoples. This matter is to be resolved.

Chemical anoll'sis of red glass trode beads 5 1 3

Assumptions

In all of the above discussions, there are both broad chemical groups and, within these. finerchemical groupings. The former are charactertzed by gross differences in specific elementalcontents. For the latter, it has been assumed that a single. 'real' chemical group is definable bysimilarities in the concentrations of the glass matrix-forming elements that are better than'10-20o/o relative'. It has also been assumed that such groupings come from either a single batchof glass. or from batches of glass made with similar proportions of the same constituentcomponents, over a relatively short period of t ime.

Finally, with the exception of the Ashuapmushuan beads, since most glass beads come fromburials. it is assumed that the time gap between the acquisit ion of beads by people in thecommunity and their interment in the ground is relatively short. Like other grave goods, glassbeads appear not to have constituted personal possessions but, rather, 'mortuary gifts' offered byothers in the community at the time of an individual's interment. Thus, they appear to have beenacquired and then subsequently removed from circulation over a relatively brief span of time.From the interregional, comparative bead groupings displayed in Table 5, this 'relatively briefspan of t ime' probably falls'within a period of approximately l0-40 years, sometimes extendingto 60 years.

C O N C L U S I O N S

Neutron activation analysis of 221 l Tth-century opaque red glass beads from New and OldWorld contexts has allowed their segregation into four discrete broad chemical groups, whichappear to be relevant to both the temporal context of the beads and to the type of beadmanufacture. The first two groups, which consist of Sn-rich beads, include all but five of thebeads from Seneca and Petun sites dating before 1655, as well as all of those from an early l7th-century Amsterdam beadmaking house, suggesting the possibility of a Dutch source for theseearly trade beads in New World native sites. While the two groups do not definitivelydiscriminate among the cored and uncored, round, circular and tubular red beads used duringthis period. cored beads tend to exhibit lower Cu and Sn contents than do uncored tubular beads.Uncored tubular beads, which are sparse in the Seneca sample, account for nearly the entire setof red beads from the Petun sites, all of which were occupied during the second quarter of thecentury.

Nevertheless. the fact that Dutch and New York (Seneca) beads are represented, albeit in lowfrequency, among these higher Cu tubular beads suggests that while the uncored tubular beadsmay have been of Dutch manufacture, they may have been obtained through some alternativeroute, rather than through the primary Dutch trade depots at Fort Nassau and Fort Orange.

The second two chemical groups of beads include all but five beads pertaining to Seneca sitesthat were occupied between 1655 and 1690, as well as the entire sample of beads from the late17th-century Algonquian site of Ashuapmushuan. The source(s) of these later beads is less clear,although striking differences in the type of red beads in use and in their chemical compositionare apparent. Most importantly, the Sn content declines dramatically in both types-uncoredround and circular beads, and cored tubular beads. These new cored tubular beads, found only atlater Seneca sites, were apparently opacified with a combination of Sn and Sb. Uncored roundand circular beads from post-1655 Seneca sites and from the Ashuapmushuan site in Quebec,however, look altogether different from any of the other chemical groups identified here. Low inboth Sn and Sb, it is not clear what the opacifying agent was. Given their divergence fromprevious red bead chemistries and the fact that the English wrested control of Fort Orange from

514 M. L. Sempowski et al.

the Dutch in 1664 and enforced restrictions on Dutch trade, it is suggested that these beads maynot have been of Dutch manufacture. Given their presence on the Quebec site as well, one mighthypothesize a French origin for the beads-either directly through French traders or via nativegroups in French Canada.

In any event, these findings regarding low levels of Cu and Sn in the later sets of beads has ledto a rejection of early speculation that a combination of Sn and Cu was technologically necessaryto form the redwood colour in glass, as a part of a sophisticated oxidation-reduction process.They do, however, open the door to speculation about the technological beginnings of Europeanredwood glass arising from the remelting of white Sn-rich glass with turquoise blue Cu-richglass.

rhis paper is dedicated ro rhe memo., ",,":::"):,

#:::::;:1r"nks go to F Neub, or the Department orMetallurgy and Materials Science, University of Toronto, for the use of his departmental ultrasonic cleaner, and to the

Rochester Museum & Science Center for the loan of the bead samples. This research was made possrble by a Social

Sciences and Humanities Rebearch Council grant to R.G.V.H., and was, initially, partially subsidised by an infrastructure

grant to the SLOWPOKE Reactor Facility of the University of Toronto from the Natural Sciences and Engineering

Research Council of Canada.

R E F E R E N C E S

Eccles, W., 1959, Frontenac, the courtier governor,97, McClelland and Stewart. Toronto.

Freestone, I. C., 1987, Composition and microstructure of early opaque red glass, in Earlt 'r ' i treius materials(eds. M. Bimson and I. C. Freestone). 151-64, British Museum Occasional Paper,56, London.

Gratuze, B.. Souher, I., Barrandon, J. N., and Foy, D., l992,De I'origine du cobalt dans les velres. Revue d'arch4omdtrie,

1 6 . 9 7 - 1 0 8 .Gra tuze ,B . ,Sou l ie r ,L ,Bar randon ,J .N . ,andFoy .D . , l 995 ,Theor ig ino f coba l tb luep igments inFrenchg lass f rgmthe

thirteenth to the eighteenth century, tn Trade and discot,en': tlrc scienti.fir' sttrdt, o.f ortefacts Ji'om post-medievol

Etrrope and ber-ond (eds. D. R. Hook and D. R. M. Gaimster), British Museum Occasional Paper, 109, 123-33.

Gratuze, B., Soulier, I., Blet, M., and Vallauri, L., 1996, De l 'ongine du cobalt: du verres )r la c6ramique, Rei'uc

d'ttrch6omdtrie. 20. 17 -94.

Hancock. R. G. V., Aufreiter, S., and Kenyon. I., 1991 , European white glass trade beads as chronological markers,

Materials issLtes in art and archaeologt', Materials Research Society, Pittsburgh, USA,462, 181-91.

Hancock, R. G. V., Chafe, A., and Kenyon. 1.. 1994, Neutron activation analysis of sixteenth and seventeenth century

European blue glass trade beads from the eastern Great Lakes Region of North America, Arc:haenrnetry', 36, 253-66.

Hancock, R. G. V., Aufreiter, S.. Kenyon. I., and Latta, M., 1999, White glass beads from the Auger Site, southern

Ontario, Canada, Journal oJ' Archaeological Science, 26, 901 -12.

Hancock, R. G. V., Aufreiter, S., Moreau J.-F.. and Kenyon. I., 1996, Chemical chronology of turquoise blue glass trade

beads from the lac-saint-Jean region of Qu6bec, in Archaeologic'ol c'hemistn,: organic, inorgunic and biochemical

anal,-sis (ed. M. V. Orna), 23-36, American Chemical Socrety, Washington, DC.

Hancock, R. G. V., McKechnie, J., Aufreiter, S., Karklins, K.. Kapches, M., Sempowski. M., Moreau, J.-F., and Kenyon.

I., 2000, The non-destructive analysis of European cobalt blue glass trade beads, Joumal of Radioanalytical and

NLtc'leur Chemistry, 244(3), 561 -13.

Houghton, F., 1912, The Seneca Nation fiom 16-55- 1 687 . Bulletin of the Bu.ffolo Soc'ieh' of Natural Scienc'es, X(2).

Houghton, F.. 1922, The archaeology of the Genesee country. Researches cmd Transac'tions of the New York State

Arc h o e o l o g i c' a l As.s o c i at i on, lll( l I ), 3 9 - 66.

Huey, P. R., 1988, Aspects o.f continuitt, uncl change in c:olonittl Dutc'h nnterial culture at Fort Orange, 1624-1664,

Ph.D. disserlation, The Depafiment of American Civil ization, Graduate School of Arts and Sciences, University of

Pennsylvania, Philadelphia.Karklins, K., 1985a, Guide to the description and classification of glass beads, in Glcss beads,8-5-123, Studies in

urc'ltaeologt', architecture ond histon', National Historic Parks and Sites Branch, Parks Canada, Environment

Canada.

Chemical analysis of red glass trade beads 515

Karklins, K., 1985b, Early Amsterdam trade beads, Ontament,9(2),36-41.Kenyon. I., 1986, Sagard's Res.sade Rouge of 1624,tn Studies in southwestem Ontario archaeologr- (ed. W. A. Fox),

53-9, London Chapter of the Ontario Archaeological Society, London, Ontario.Kenyon. I., Hancock, R. G. V., and Aufreiter, S., 1995, Neutron activation analysis of A.D. 1660-1930 European copper-

coloured blue glass trade beads from Ontario, Canada, Archaeometry",37,323-31 .Kidd, K. 8., and Kidd, M. A., 1970, A classil ication system for glass beads fbr the use of field archaeologists, Canadian

Historic Sites: Occasional Papers in Archaeolo{v- and History,1, 45-89.Kupp. J., 1974, Aspects of New York Dutch trade underthe English. New York Quorterlt ', LVIII(2).Moreau, J.-F.,1994. Des perles de la 'protohistoire' au Saguenay-Lac-Saint-Jean, Rec'herches amdrindiennes ou Qudber',

24 (1 -2 \ .31 -48 .Moreau, J.-F.. Hancock, R. G. V., Aufreiter, S., and Kenyon, I., 199'7. Taphonomical and chronological studies of a

concentration of European glass trade beads from Ashuapmushuan, central Quebec (Canada), in Proc:eedings of theseventh Nordic conJerence on the oppliccttions oJ'science in orc'hoeolog,r'(eds. H. Junger and M. Lavento), 173-81,Finska Fornminnesfcjreningen, Iskos, Finland.

Sagard. G., 1865. Le grand rorage du pa,-s des Hurons,2 vols, Paris, Tross.Sagard. G., 1866. Histoire du Canada,4 vols, Paris, Tross.Sempowski, M., Nohe, A. W., Moreau, J.-F., Kenyon,I., Karklins, K., Aufreiter, S.. and Hancock. R. G. V., 2000, On the

transition frorn tin-rich to antirnony-rich European white soda-glass trade beads in northeastern North America,J ournal of Rodioortulytical and Nuc'l e at' Chemistn', 244(3), 559-66.

Sempowski, M. L., and Saunders, L. P., 2001. Dutch Hollow and Factory Hollow: the advent of Dutch trade among the

Seneca, Charles F. Wroy Series in Senecq archaeologt', vol. 3, Research Records #24, Rochester Museum & ScienceCenter. Rochester, New York.

Sou l ie r ,L .Gra tuze ,B . ,andBar randon .J .N . , l 996 ,Theong ino f coba l tb luep ig rnen ts inFrenchg lass f romtheBronzeAge to the eighteenth century, inArchaeontetn ' 1994 teds. S. Demirc i , A. M. Ozer and G. D. Summers). 133-40,Ankara, Turkey.

Thwaites, R. G., ed.. 1896-1901 ,The Jesuit relatiorts und all ied documents, T3 r,ols, The Burrows Bros. Co., Cleveland,Ohio. {

Trigger B. G.. 1987, The children o.f Aatuentsic. A histon'oJ'the Hurr.tn people to 1660, Montreal. McGill-Queen'sUniversity Press.

Wray, C. F., 1973, Manual for Seneca archaeologt', Cultures Primitive, Honeoye Falls, New York.Wray, C. F.. and Schoff. H. L., 1953, A preliminary report on the Seneca sequence in western New York (1550-1687).

P e n n st' I v cut i a A rc h u e o I ogi.lr, XXII( 2), 5 3 -63.

Wray. C. F.. Sempowski, M. L., and Saunders, L. P., 1991. Tram and Cameron: two early contact era Seneca sites.Chorles F. Wrat'Series in Senecet arc:haeologt, vol.2, Research Records #2 1 (ed. C. F. Hayes III), RochesterMuseum & Science Center, Rochester, New York.

Wray. C. F., Sempowski. M. L.. Saunders, L. P.. and Cervone. G. C., 1987. The Adams and Culbertson sites, Charles F.Wrat' Series in Seneca archaeologt', vol. 1, Research Records #19 (ed. C. F. Hayes III), Rochester Museum & ScienceCenter. Rochester. New York.