Some observations on two early Cape florilegia

A number of early Cape florilegia and codices exist in libraries in Europe and South Africa. Four of these florilegia are closely related and are housed in the Brenthurst Library, Johannesburg, the Botanical Research Institute, Pretoria, the Bodleian Library, Oxford and the Rijksherbarium, Leiden. The first two are compared and discussed in detail in this article. Arising from this comparison, a new interpretation of the interrelationship and origins of the four florilegia is proposed. The key volume is the florilegium in the Botanical Research Institute, Pretoria.


INTRODUCTION
The Cape flora has for three centuries excited the interest o f botanists.W ith the rem arkable increase in th e exploration o f the w orld from the fifteenth century, new areas w ith fascinating plants and anim als becam e know n to the educated and knowledgehungry w orld o f E urope.The C ape o f G o o d H ope became a vital stopping-over po in t for the m any ships going to the E ast in search o f spices and riches.It was thus inevitable th a t plants growing at the Cape found th eir way back to E urope.Soon the dem and for these unusual plants, both to grow and to possess as dried specim ens o r paintings, increased considerably.
D uring the governorship o f Sim on van der Stel from 1679 to 1699, there began a period o f considerable ex ploration and expansion.This was accom panied by a significant increase in the scientific knowledge o f the indigenous flora and fauna with Van der Stel him self as the w orthy patro n .It is known th a t he com m issioned the artist C laudius to record by m eans o f sketches the n atu ral history o f the area and also the gardeners, O ldenland and later H artog, to build up " one o f the m ost beautiful and curious gardens I have ever seen" (T achard, 1686).
Illustrated books on the C ape flora were not available at the tim e an d so collections o f paintings, florilegia or codices o f anim als were also included, were produced fo r influential p atro n s o f n atural history.A m ong these florilegia o r codices were the Codex W itsenii, C odex B entingiana, Codex Com ptoniana, D olneus's Florilegium , Van der Stel's own Collection and the C odex accom panying the official report o f his E xpedition to N am aq u alan d in 1685/6. G unn & C odd (1980) give a fine overview o f early Cape botanical history in which they discuss these works and their significance.
Examples o f early C ape florilegia and codices exist today in various institutes and libraries in Europe and South A frica, nam ely: The w orks m entioned in this paper will be referred to by the abbreviations given above for the institutes and libraries where they are housed.M uch has been published ab out these w orks by W aterhouse (1932( B arnard (1947( ), Sm ith (1952)), Jessop (1965), Edw ards (1965( ), K ennedy (1967( ), M acnae & D avidson (1969( ), G unn & D u Plessis (1978)), W aterhouse (1979) and G u n n & C odd (1980).
The florilegia in the first four institutes m entioned above are o f particular interest to me because o f their close relationship, one o f them being in the library o f the Botanical Research Institute, Pretoria.Jessop (1965)  It was not possible for me to ascertain the exact binding sequence as the original book had recently been restored.They do, however, m ention th a t forty-nine leaves have no fold-m arks an d fortynine have fold-m arks which indicate th a t these folded leaves were painted on at the C ape, folded and then despatched to Europe.They also m ention the extra pieces o f paper which had been glued into these folded leaves, when binding took place, in order to bring all the pages up to the sam e size.Some o f the paintings on the larger sheets had in tu rn been cut during binding.This is m ost noticeable on the coloured frontispiece and plates BFC 30 and 42.
BRI is b ound into gatherings o f 5 or 6 sheets with one gathering o f 7 sheets.These sheets are folded giving gatherings o f 10, 11, 12 or 13 pages w ith three instances o f single pages having been tipped in during binding.Jessop (1965) points out th a t the sam e cutting o f paintings occurred in BRI and feels th at the paintings m ust have been executed before the book was bound.The cut paintings occur only on BRI 8, 68, 86 and 109 to any extent, but far less th an in BFC.O ne's interp retatio n o f the term " binding" is, to my m ind, im portant.I feel th a t the book was m ade up with a softish cover, as is found in SAPL, to form a w orking volum e into which the paintings could be executed.A t a later stage the volum e was slightly trim m ed when the soft cover was replaced by a proper vellum binding for placing in some person's library.The 68 blank pages at the end o f BRI give a very strong indication th a t the w ater-colours were painted into the book.A nother point which supports this view is the statem ent by T achard (1686) th at Claudius " had already com pleted two thick volumes o f divers plants, painted from n atu re" .Churchill (1935) states th a t the Strassburg Lily with 4W R was first used in 1636 and is o f G erm an origin.Jaffe (1930) andH eaw ood (1950) say it was made at the paperm ill o f one W endelin Richel (Riehel), which began production in 1583 w ithout counterm arks.This w aterm ark becam e associated with quality paper and was m uch copied in Europe.Jaffe also states th a t the co u nterm ark, IHS sur m ounted by a cross, " /« Hoc S ig n o '\ was first used in the Lom bardy area in Italy from 1481-1580 before being taken up in the L othringen area o f G erm any.The version with DYSVLI was one o f the many counterm arks used by J. Villedary (Vildary) whose mills produced paper over a period o f 150 years from 1658-1812.The counterm ark with RM is unknow n, but noted by H eaw ood (1950).

The oth er florilegia in
It is not possible, using currently available references, to put any date to the m anufacture of the paper in BRI.Jessop (1965)
Erica cerinthoides, but did not use any BFC paintings for these plates.T hus it can be assum ed th at he did not possess any B FC paintings at the time.This means th at he m ust have received the paintings from the C ape betw een 1678 and his death early in 1697.

THE PAINTINGS
The BFC volum e contains w ater-colour paintings The last example is illustrated in Fig. 2 where a glance will show the m ore natural p ro p o rtio n s o f the BRI painting.This feature w ould indicate th a t the BRI paintings could n o t have been copies from BFC an d th a t the reverse is the case.The act o f copying is always fraught with the possibility o f slips, m isinterpretation or plain reinterpretation by the copier.This is evident in the differences th a t occur between some o f the BRI and BFC paintings.In copying the artist has produced extra curliness or waviness into the organs th a t he was copying.This is clearly seen in the leaves o f Empodium plicatum shown in Fig. 3  As I am interested in ericas, I was particularly struck by the differences between the tw o sim ilar renderings o f Erica cerinthoides.In BRI 43 the leaves are show n in distinct w horls an d are themselves depicted trigonous whereas in BFC 90 the leaves are random ly scattered and executed rath er poorly and haphazardly by the single stroke o f a brush.One can also see th at the copier in BFC could not easily in terpret the BRI flowers which m ust have been som ew hat passé when painted.
A noth er interesting and im p o rtan t feature, which was also noted by Jessop (1965), is th at there are several paintings in BRI which have faint pencil outlines still rem aining on the pages.These  (1978) and by Jessop (1965) th at some o f the paintings could have been executed from plants cultivated in Europe.As many o f the plants in BFC and BRI are geophytes they could easily have been grown in Europe at the time.However, one or tw o features point to a wild origin for the plants, certainly o f the G roup A species.These include the dam age to the leaves caused by insects and other anim als, and so accurately portrayed by the artist.The best example of this feature is BRI 42, BFC 35 o f Babiana tubiflora which has its leaves alm ost com pletely chewed off by some grazing anim al.As G unn & D u Plessis (1978) point out " This is a very clear indication th at the painting was m ade from a plant which grew wild at the Cape and not from a cultivated p lan t" .F urther examples o f this type may be found in BRI 45, Wachendorfia paniculata; BRI 52, Ornithogalum thyrsoides and BRI 92, Ixia paniculata.Also, the natural dim ensions o f the vegetative parts o f the plants suggest th at the subject was a wild plant rather than one grow n under glasshouse conditions in Europe.
In this discussion o f the paintings in BFC and BRI it is w orth m entioning th at in BRI there are various pieces o f plant debris lodged in betw een the pages.In the fold with BRI 131, on which the unnum bered painting o f Erica cerinthoides occurs, there are three leaves o f Erica cerinthoides.These leaves could either have become lodged in the fold when the species was being painted or at a later stage when the owner o f the volum e was perhaps com paring a specimen with the painting.

NUM BERING
The paintings in all o f the four sim ilar florilegia are num bered, according to G u n n & D u Plessis (1978).These num bers provide an im portant feature for com parison.The im portant volum e is BRI.In this volume nearly all the paintings are num bered near the base of each plant in a consecutive sequence up to 142. Strangely the first painting is not actually num bered.N um ber 142 is followed by 10 un num bered paintings and then the last two in the set num bered as 143 an d 155.The sequence o f sheets in the binding gatherings has not been interrupted.F our pages have, however, been cut out of BRI leaving gaps in the num bers, viz.BRI 13, 14 and 15;27;50;90 and 91. As G unn & D u Plessis (1978) state, com parison with BFC gives the identity of two o f these missing paintings, because copies of BRI 27 and 90 occur in BFC.An additional three pages have been cut out o f BRI near the end o f the paintings, but, as no num bers are missing in the sequence, one may assume th at these pages were removed before the num bering was done as stated by Jessop (1965).These pages may or may not have had paintings.Despite the anom alies in the num bering tow ards the end o f the paintings htere should have been a total o f 155 paintings, but with the loss o f seven there now rem ains the total o f 148 The first set o f num bers up to 98 is w ritten in a very neat small w riting with black ink while the num bers 99-143 are w ritten in a different slightly larger writing and paler ink.N um ber 155 is w ritten in another handw riting and num ber 57 in yet another.In several cases already m entioned, two com pletely different styles o f painting occur on the same page and also on the same sheet o f paper but separated due to the binding.These are all num bered in the sequence.Why there are two different m ain handw ritings in the num bering and w ho did them cannot be answered.The num bers m ight have been done by the artists.Why there is the batch o f 10 unnum bered paintings near the end is also a mystery.The fact th a t the paintings were num bered consecutively regardless o f the painting styles shows th a t the num bering m ust have been done directly into the volum e and m ust therefore be a series exclusive to BRI.
BFC has 85 pages num bered consecutively from 2-86.These num bers have been w ritten all in the same handw riting in the to p right-hand corner o f the recto o f each page.These num bers refer to the pages and not the paintings as on 14 pages there are two, three o r four paintings.G u n n & D u Plessis (1978) give references to these 'Fo lio ' num bers.In BFC there are 35 paintings bearing num bers near the base of the plant depicted an d these num bers are the same as those on the BRI paintings o f the sam e species.All the paintings bearing these num bers in BFC belong to the superior G ro u p A, while all the G ro u p B paintings in BFC are unnum bered (cf.Fig. 2).
G unn & D u Plessis (1978) miss the real significance of this very im p o rta n t p o in t which gives additional proof th at the BFC G ro u p A paintings m ust have been copied from BRI. N one o f the 31 paintings in BFC belonging to G ro u p B and shared with BRI is numbered.This suggests th a t the BFC G ro u p B paintings were copied from some originals in another volume w ithout any num bering and th a t the G ro u p A paintings were later copied from BRI together with the noting o f the BRI num bers.
Two paintings w ith num bers stand out as unusual.BRI 63, Polygala bracteolata, a G ro u p A painting, occurs in BFC as a G ro u p B painting (cf.Fig. 3).If the copying o f the G ro u p A paintings had been so good why had this species been copied so poorly.The other unusual painting is BFC 13, Hessea cinnamomea, G ro u p A. In B FC it bears the num ber d istin ctly w ritten as 105, but BRI 105 is o f a G ro u p B leguminous species.H. cinnamomea in BRI is 108, but the 8 is ra th e r indistinctly w ritten and at a glance could well be m istaken for a 5 which is most probably w hat the copier did.The LD " volum e" is in fact a set of loose paintings o f inferior quality which were acquired by D. van Rooyen at two auctions in 1778and 1779. G unn & Du Plessis (1978) suggest a possible link between some o f these paintings and the collections o f Seba.
There occurs in the BRI volume with the num bering on only 17 paintings in the first 96, a plus sign.In some cases this plus sign looks as though it was done by the num berer, in others by B urm an (cf.Fig. 4).All o f the 17 paintings belong to G roup A. The significance of this sign is as yet not understood.

COMMON NAMES
BFC and BRI have a num ber of paintings accom panied by a com m on nam e in D utch (cf.Fig. 3) or occasionally a Latin polynom ial.They are all w ritten in the same, but very different, handw riting in each volume.These com m on names were given to nearly all of the G roup B paintings but a few were given in BRI to G roup A paintings by the same person (cf.The handw riting in BRI is very distinctive and could well be the same as th a t used to write the notes accom paying the anim al paintings in the C odex W itsenii, SA M ; cf.FO L. 160 reproduced by B arnard (1947).This handw riting is not the same as th a t occurring on the botanical paintings which, according to Sm ith (1952) quoting a form er C hief A rchivist, G raham Botha, is identical with th a t in IPA .T he handw riting o f the notes in SAPL is also o f a sim ilar style.This could suggest th a t the artist w rote the nam es in BRI and the notes in the SAM anim als, which are regarded as having been copied a t the Cape for N icolaas W itsen some time before 1692 (B arnard 1947).
A num ber o f the com m on names differs slightly between B FC and BRI, m ost being differences in spelling which could be attributed to the hom e language o f the copier, e.g.m iddags/m iddaghs; bloem /blom ; R ingel/R engel; -aanse/-aense; sterre/ starre.Some peculiar differences are noted here:

ANNOTATIONS
The BRI volum e was a t one time in the possession o f Johannes B urm an as it was inscribed by him on 3 August 1755 [cf.fig. 1 o f Jessop (1965)].Burm an also annotated every painting in BRI.M ost o f his annotations were p robably done a b o u t the same time as the same d ark ink an d style o f w riting was used by him.This was done in the volum e as is evidenced by the blotting o f the ink on the verso o f the preceding pages.He quotes m any times Breynes's Prodrom i o f 1739 and his own R ariorum A fricanarum Plantarum o f 1738/39 using Latin polynom ials.However, on BRI 112, Antholyza ringens, he cites Linnaeus's Species P lantarum o f 1753 and quotes his description.Binom ials were added by Burm an using a finer pen.A dditional an n o tatio n s were also added in a larger clum sier handw riting using paler ink.

PUBLISHED WORKS
Paintings from b o th B FC and BRI are know n to have been used in published w orks, namely, Breynes's P rodrom i (1739) and B urm an's R ariorum A fricanarum P lan taru m (1738/39).M ention has already been m ade o f these by Jessop (1965) andby G unn &D u Plessis (1978).H ow ever some extra observations n o t no ted by them throw a different light on the relationships o f the florilegia and the above publications.
The younger Breyne undoubtedly used the BFC paintings as the originals from which 15 o f the engravings in the P rodrom i o f 1739 were copied.These paintings occur in b o th BFC and BRI but a very careful com parison o f the w ater-colours and the engravings shows th a t the engravings were m ade from BFC.All o f these originals fall into the G ro u p A paintings.Seven engravings can be linked to w ater-colours which occur only in BFC and these are all G ro u p B paintings.
O f significance are three engravings which can be linked to w ater-colours o f G roup A found only in BRI.Tab.VII fig. 2  Wachendorfia paniculata which is taken from BRI 45 and is illustrated in the present article in Fig. 4. O f p articular interest is the third example, Monsonia speciosa on Tab.X X I, fig.2, which is taken from BRI 22 (cf.Fig. 4 in the present article).On the sam e T ab., fig. 1 is o f Senecio cymbalarifolius which is represented in both florilegia, G roup A paintings, and in the text is cited as being " ex Flora nostra Capensis" .But in the text for Monsonia the Breynes state " w Flora nostra Capensi" .This w ater colour does not exist in BFC.From their statem ent it w ould appear th at the Breynes had had access to a copy o f this species, which had not been bo u n d into BFC, or even access to BRI.If the latter were the case why then had the Breynes not reproduced m ore o f the superb w ater-colours.
J. Breyne had collected together the paintings for B FC before his death in 1697.His son, J. P. Breyne, when m entioning the paintings in the Prodrom i did not give exact details o f their origin other th an th at they cam e from the Cape, e.g.Tab.X II, fig. 1 B urm an's R ariorum A fricanarum P lan taru m con tains descriptions and engravings o f some C ape plants.B urm an attributes 92 o f them to the Codex W itsenii, 34 to H erbarium W itsenianum and 33 to the C ollection or Codex Sim on van der Stel.O f all these figured plants six can be identified as being alm ost identical to w ater-colours in BRI.In the text accom panying the plates he attributes five o f the six species to the Collect, van der Stel.e.g.T ab.X II, fig.2, " atque haec in Collect, van der S tel eleganter depicta m ihi obvenit, unde hanc exhibem us".All o f these species depicted are G roup B paintings.The rem aining 28 species attrib u ted to the Collect, van der Stel do not occur in BRI.Therefore it m ust be assum ed th a t B urm an had all these plants reproduced as engravings from a volum e which he knew was the Collect, van der Stel an d th at the five G ro u p B copied w ater colours in BRI came from th a t source.As stated by Jessop (1965) and by G u n n & C odd (1980) this volum e is n o t traceable.
The rem aining BRI w ater-colour depicted in B urm an's w ork is o f Oxalis purpurea on T ab.X X V II, fig. 3 and is attrib u ted to the Codex W itsenii.This is the figure th a t Jessop was concerned a b o u t because o f the hairiness an d stam ens which B urm an had added.Reference to the text shows th a t B urm an referred to other w orks in which the species was m entioned, namely those o f Com m elin who described his species as glabrous, o f Breyne as hirsute (in fact only the calyx) and o f H erm an also as hairy an d had thus ad ap ted his figure accordingly.There is, o f course, the possibility th at the copier for BRI had merely om itted the hairs and stam ens, which were present in the original.
In Fig. 5 o f the present article show ing Pelargonium cucullatum, the centre illustration is taken from B urm an's w ork, Tab.XXXV.He cites in the text " & ex Collect.D. van der S te l hanc publici juris fa c im u s" .O n the left in Fig. 5 there is the same species as depicted in BRI 103 and on the right is B FC 54.T he similarity between the BRI w ater-colour and the B urm an engraving is obvious.The BFC w ater-colour is rath er far rem oved, b u t can be seen to bear som e resemblance.Both w ater-colours belong to the G ro u p B copied paintings.It has been show n earlier in this article th at the BRI copies are probably tru e r copies o f the originals th an the BFC copies.One w ould then assum e th a t the BR I and BFC paintings were copied from the same original which was in the Collect, van der Stel.The oth er two species o f Pelargonium illustrated in this plate were copied from the Codex W itsenii and are alm ost identical to the w ater-colours attrib u ted to C laudius in SA M , IPA and T C D .
One plate, Tab.LXXV, in B urm an's w ork is o f special interest.In it are depicted Crassula capensis (fig.4) and an unidentifiable com posite (fig.5).The form er is present in b o th BFC and BRI, but the latter is only in BFC.In the text B urm an states for the Crassula "Fructus nec semina adpicta sunt in Collect.D. van die S tel, unde hanc cum subsequente producim u s'\ the subsequent figure being the com posite.This would indicate th at the Breynes had had their painting copies from the Collect, van der Stel.
B arnard (1947) gives a detailed description o f SAM which he says is p art o f the Codex W itsenii particularly as it is autographed by W itsen.The C odex contains 12 w ater-colours which B urm an figured and attrib u ted to the C odex W itsenii.These engravings were taken from SAM an d not from the alm ost identical paintings in IPA , SA PL o r T C D , as a detailed com parison o f the paintings and engravings shows a closer m atch w ith SAM .M any o f the plates which B urm an attrib u tes to the Codex W itsenii can be traced to th a t very fine Codex, IPA, in Johannesburg.This has been discussed by M acnae & D avidson (1969).

CONCLUSIONS
The four early florilegia housed in Libraries in Pretoria, Johannesburg, O xford an d Leiden consist o f the same basic set o f w ater-colour paintings of Cape plants.These paintings can be divided into two distinct groups, A and B, on the style o f painting and the quality o f detail.The key volum e is the florilegium, BRI, housed in P retoria.
In BRI the quality o f the G ro u p A paintings is outstanding, as the paintings have m ore natural proportions, in some cases contain m ore a n d better details and were executed with good quality paint.They m ust be regarded as originals.The G roup B paintings are all reasonable copies taken from another or, perhaps, several sources.The paper on which the paintings were executed is all o f sim ilar m ake and quality and is unfolded.All the paintings were executed at the Cape directly into the volum e, as is indicated by the occurrence o f G ro u p s A and B random ly distributed through the volum e, some on the same sheet and others even the same page.The volume was later properly bou n d in vellum.The paintings were nearly all num bered consecutively giving a series o f num bers relevant only to BRI.
In BFC the G ro u p A paintings are o f poorer quality with sometimes fewer details depicted, they have altered unnatural proportions and were painted with a poorer quality w ater-colour paint.They all bear a num ber which corresponds to th a t in B R I, but are random ly arranged at the beginning o f BFC.These paintings are undoubtedly copies o f some o f the G roup A paintings in BRI, and are done on unfolded paper.The G roup B paintings o f which there are many m ore th an in BRI, are likewise copies taken from a sim ilar source as BRI.These are all done on folded paper which indicates th a t the copying was done at the Cape.The G roup A and B paintings were copied separately on several different types o f paper and then bound into a volum e as late as 1724, with G roup A paintings first and G ro u p B's second.At this stage the pages or folios, n o t the paintings, m ust have been num bered consecutively.
The volumes, O X F and LD , are both sets o f inferior copies with their num bering related to the folio num bers o f BFC.This indicates th a t they m ust have been copied from BFC after 1724.The copies in LD were acquired at different tim es, the G roup A paintings in 1779 and the G roup B paintings in 1778, possibly from the estate o f Seba according to G unn & D u Plessis (1978).
Burm an used some G ro u p B paintings from which engravings were m ade and published in his R ariorum A fricanarum P lantarum o f 1738/39.He stated th at they came from the collection o f Simon van der Stel.This could indicate th a t all the G roup B paintings were copies for BRI a n d BFC from one o f the volumes o f paintings owned by V an der Stel and now untraceable.The m ain questions th a t rem ain are-who were the artists and when was the first florilegium painted?The first question will probably rem ain as a point for speculation an d rem ain unansw ered for ever, as none of the artists active at the Cape in its early days ever signed a copy o f his w ork.W hen the BRI florilegium was executed can be roughly deduced from certain facts.Jak o b Breyne m ust have acquired his collection o f paintings before he died in 1697.Thus for BFC G ro u p A paintings to have been copied from BRI, BRI m ust have been in existence before 1697.As Breyne did n o t use any o f the BFC paintings to illustrate the Cape plants depicted in his C enturia o f 1678, it can be assum ed th at he acquired BFC after 1678.G u n n & D u Plessis (1978) m ention several artists w ho were active at the Cape from the mid-1680's to the mid-1690's.It is reasonable to assume th at BRI was painted during th a t period.
As G unn & C odd (1980) state, botanists in South Africa m ust be grateful to the B renthurst Press for publishing the com plete BFC florilegium in colour.This statem ent I certainly endorse.A lready available as reproductions are the superb facsimile edition in colour of T C D (W aterhouse, 1979), the sepia reproduc tion o f IPA (K ennedy, 1967) and the rather poor black and white reproduction o f the SAM paintings (Bar nard, 1947), all o f which give botanists an idea o f the paintings in those volum es and som ething with which to make com parisons.As yet BRI, SAPL, O X F and LD have not been reproduced in any form to make them generally available to researchers.G unn & Codd (1980) also po in t to the possibility o f the existence of " undiscovered" m anuscripts and volum es in libraries and archives in E urope, particularly in the rich collec tions at Leiden.It is hoped th a t this article will add to the increasing literature on early C ape florilegia and that at some time in the future new inform ation will come to light th a t will solve som e o f the unansw ered problems.

ACKNOWLEDGMENTS
Sincere thanks are due to the B renthurst Library in Johannesburg for allow ing exam ination o f the original copy o f the Breynes's F lo ra Capensis.This also applies to the A fricana M useum in Johannesburg and to the South A frican M useum and South A frican Public L ibrary in C ape Tow n for perm itting exam ination o f the original codices in their collections.
Useful discussions were held w ith D r L. E. C odd and with Miss M. D. G unn. M rs A. R om anow ski is thanked for her reproduction o f the w ater-colour paintings.

U lT T R E K S E L 'n A antal vroeë Kaapse florilegium s en kodekse bestaan in biblioteke in
published a detailed account o f this florilegium, b u t unfortunately at the tim e did n o t know o f the existence o f the other three florilegia.G unn & D u Plessis (1978) edited and w rote the in troduction to the so-called F lora C apensis o f Jak o b and Jo h an n Philipp Breyne, housed in the B renthurst Library and which was reproduced in toto by the B renthurst Press, Johannesburg.They were unable to give a detailed com parison o f the four florilegia.Resulting from my close exam ination o f BFC and BRI while w riting a review o f the BFC reproduction for B othalia (Oliver, 1980), I have been able to extend the published notes and observations on the tw o florilegia in South A frica.I have n o t had the o p p o rtunity to exam ine the florilegia in O xford and Leiden and accept the statem ents m ade by G unn & D u Plessis (1978), who have consulted them .In the discussions which follow, reference is m ade to the num bering o f the plates (as arranged by G u n n & D u Plessis) in the B renthurst reproduction o f the Breynes's F lo ra C apensis BINDING The m ost noticeable difference between B FC and BRI is in the size o f the volumes, where the page size in B FC is 310 x 195 mm and in BRI 3 9 3 x 2 5 0 mm.This difference in size is significant as will be seen when the origins o f the tw o florilegia are discussed later on.The B renthurst reproduction o f B FC has been reduced to an even sm aller size, the title page being 230 mm long instead o f 270 m m (G u n n & Du Plessis, 1978).Reference to the list o f plates at the beginning o f BFC shows th a t 47 o f the 102 paintings had to be reduced.The binding o f BFC is full red m orocco leather w ith gold tooling including the Breynes's coat-of-arm s.The pages have been gilded.O n the title page it is stated th a t the volum e was b o u n d by the younger Breyne in 1724.BRI on the o th er hand is plainly bound in vellum, m uch like the codices SA M and SA PL ; the tooling is blind and the edges o f the pages have been spattered w ith red an d blue ink.O n the front cover there is a distinct erect capital P done in slightly faded black ink.B arnard (1947) m entions a capital Q on the front cover o f SAM .I exam ined SA M and find th at the Q slants obliquely to the right, is sm aller and is placed at the to p o f the fro n t cover.G u n n & D u Plessis m ention the binding o f the pages o f BFC, but give no details a b o u t the gatherings.
South A frica are rath er sim ilar to BRI in appearance.IPA is m ost like BRI being sim ilar in size and vellum binding.The paper is o f the same thick quality and the paintings are done on the recto o f each page w ith binding in gatherings.In SAM and SAPL the volum es are m uch sm aller, like BFC.In these two florilegia the binding o f the sheets is interesting.The leaves are folded double an d bound singly giving a lot o f 2-paged gatherings.The paintings occur on the recto o f the first page with the descriptions on the verso in SAM and the recto o f the second page in SAPL.Both G unn & D u Plessis an d Jessop give details o f the w aterm arks found in the respective florilegia and discuss the possible origin and date o f m anu facture o f the paper.In BFC there is a num ber of different types o f paper.G unn & D u Plessis did not, however, m ention the connection between w ater m arks and counterm arks presum ably due to the difficulty o f unravelling the binding sequence o f so m any varied sheets.They do m ention the similarity o f the w aterm arks in BFC, SA M and T C D .BRI, on the other hand, has very uniform paper o f only two types.Jessop gives an illustration o f the main w aterm ark, the Strassburg lily with coronet, m antling and 4W R and the m ain counterm ark, IH S/D Y SV LI (cf.left-hand pair, Fig. 1 in the present article).He noted th a t a few pages possessed a different countermark, IH S/R M .This is, in fact, the counterm ark for a slightly different w aterm ark which is more crudely produced and o f sm aller size (cf.right-hand pair, Fig. 1).The counterm ark itself is very much cruder, as well, and this suggests either an inexperienced paperm aker or paper o f earlier origin before the refinement o f the w aterm arking technique.This cruder w aterm ark occurs in paper th a t takes up two bound gatherings o f 22 pages between paintings 111 and 132.The m ain w aterm ark occurs a num ber o f times in IPA, but with different counterm arks.It also occurs in the C laudius anim al paintings in the A fricana M useum in Johannesburg, as figured by Smith (1952).It can also be seen in the binding paper o f Com m elin's H orti Medici A m stelodam ensis (1697-1701) and in the body o f the book where different m antling and the counterm ark of Pieter van der Ley occur.Both the m ain w aterm ark and counterm ark o f BRI occur once in BFC.
of 102 species o f w hich 66 occur in BRI and 36 are exclusive to BFC.Included in the last set are two paintings which are know n to have occurred in BRI, but are now m issing (see under N um bering).G unn & D u Plessis (1978) recognized two different qualities o f art-w ork and grouped the paintings into G roup A o f superior quality and G ro u p B o f poorer quality.They ascribe 37 paintings to G ro u p A and 65 to G ro u p B. I w ould, however, have placed the two non C ape paintings, BFC 60 and 69, in G ro u p A. I would also regard B FC 18, Spiloxene alba, as being a G roup A painting and BFC 42, Chasmanthe aethiopica, as G ro u p B. This leaves all the 36 Cape species exclusive to B FC belonging to the G ro u p B painting quality.I have followed the same arrange ment o f groups for convenience in this com parison.The BRI volum e contains w ater-colour paintings of 148 species o f which there are the 66 shared with BFC and 82 exclusive to BRI.In my opinion the 148 paintings can be divided into 115 being G ro u p A and 27 being G ro u p B w ith 6 paintings being impossible to place with certainty.All o f the 82 paintings exclusive to BRI are o f G ro u p A quality.The G roup A paintings, however, can be subdivided into a possible tw o o r even three different styles.The m ajority m atches the G ro u p A style paintings in BFC.A t the beginning o f BRI there is a series o f paintings, nam ely 3, 5, 6, 7, 8, 9, 10, 17, 18, 19 and 20, which are in a com pletely different style, rather vague and lacking precise details and yet not crude like some o f the G ro u p B paintings.These paintings are undoubtedly the w ork o f a different artist and strike me as being originals.Their quality and colours remind me o f the w ork o f the " second a rtist" in IPA.M ost o f the G ro u p A paintings are accurate and beautifully executed.T ow ards the end o f the series of paintings the quality o f the p ain t seems to change, the greens being m uch deeper in colour w ith a bluer tinge and the p ain t is m uch thicker.These paintings could also be the w ork o f an o th er artist.One painting, BRI 135, stands o u t as com pletely different from the rest o f the num bered plates.It is o f Crassula coccinea and is classified as a G roup B painting, but gives the strong im pression o f having been executed in the C laudius style like the paintings occurring in SAM and SAPL.The paint is o f quite a different texture and the painting lacks any perspective.The extra painting o f Erica cerinthoides, an unnum bered one occurring with BRI 132, is m entioned by Jessop as also unlike any others in style and quality o f paint.This paint is very rem iniscent of the pow dery type used in the G ro u p A paintings in BFC as it is sm udging slightly and has left an im print on the recto o f the preceding page.This feature also occurs on BRI 99 in the brow ns o f the rather large tu ber o f the Bulbine tuberosa.The differences between the G roup A and G ro u p B paintings in BRI are very m arked, particularly where the tw o styles occur on the same page, e.g.BRI 55, Lobelia pinifolia (A) and 56, Drosera cistiflora (B); BRI 71, Pelargonium longifolium (A) and 72 M oraea tricuspidata (B); BRI 75, M esem bryanthem aceae (A) and 76, Galaxia ovata (B) and 77, Dorotheanthus bellidiformis (B).In these examples it is always the G ro u p A painting which is on the left-hand side of the page and num bered first.This would indicate th at the small G ro u p A painting was done first leaving space on the page for additional species to be illustrated at a later stage which, in the above cases, turned out to be rath er crude G roup B paintings.This feature o f the tw o distinct qualities o f paintings on the sam e page is im p o rtan t in respect to the problem o f the origin o f these florilegia.G u n n & D u Plessis (1978) are o f the opinion th a t the paintings o f BFC are the originals in the set o f four very sim ilar florilegia.They based their opinion on the dating o f the paper-" The paper used in the Breynes's 'F lora C apensis' is the earliest and o f a period consistent with the possibility th a t it is the original set" .This m ight be true o f the paper, b ut n o t o f the paintings, as the following points obtained from a detailed com parison o f BFC and BRI will show.G u n n & D u Plessis (1978) state th a t the volum es in O xford and Leiden are obviously inferior copies, which opinion I m ust accept n o t having had the o p p o rtunity to exam ine them myself.The m ost noticeable difference between the B FC and BRI paintings in the G roup A series is the shortened dim ensions o f m any o f the paintings in BFC.The floral parts o f the paintings are painted the same size as those in BRI, but the vegetative parts, particularly the stems, have been reduced to be able to fit onto the smaller size o f the B FC paper, in som e cases so drastically as to m ake the painting dispropo rtio n ate.G ood exam ples o f this feature may be seen when com parisons are m ade o f BFC 19 and BRI 125, Spiloxene capensis; B FC 20 and BRI 126, M oraea aristata; BFC 40 and BRI 109, Gladiolus maculatus and BFC 38 and BRI 68, Gladiolus carneus.
Fig.2also illustrates an o th er im p o rtan t feature, namely the lack in BFC o f certain details present in BRI paintings.In the BRI paintings o f Gladiolus carneus there are an additional flower bud, term inal bracts and a leaf.There is also a n o th er very im p o rtan t and significant feature ab o u t this BFC painting, and th at is the lack o f paint in the region o f the ovary o f the basal flower, where the artist forgot to fill in the colour.This also occurs in BFC 39, BRI 122, Gladiolus hyalinus.F u rth er exam ples o f parts o f the plants being left out in the BFC paintings can be found in paintings which lack roots, hairs, flowers, rtr e ra n i t J ? a f r i.* 147 Fïg.3.-Polygala bracteolata L. Left, painting number 63 from the BRI florilegium; centre top, painting number 81 from the BFC florilegium (Brenthurst reproduction), both x 0 ,5 .Empodium plicatum (L.f.) Bak.Right, painting number 97 from the BRI florilegium, x 0 ,5 ; centre bottom, painting number 62 from the BFC florilegium (Brenthurst reproduction), x 0 ,7 0 .
are o f additio n al parts o f the plants which the artist m ust have decided not to use.Two clear exam ples are BRI 12, Adenandra villosa, and BRI 109, Gladiolus maculatus.This feature gives a strong indication th at the paintings are originals and not copies.The strangest anom aly found in the com parison o f the paintings occurs in the paintings o f Sparaxis bulbifera, BRI 102, BFC 32, both G ro u p B paintings.The paintings have their lower halves reversed.Reversing o f paintings is o f course com m only encountered in engravings used for printing plates.From the above com parison o f the paintings it is my opinion th at the G ro u p A paintings in BRI are the originals and th at the BFC paintings were all copied from BRI.A n exam ination o f the paints used lends additional weight to this view.In BRI the w ater-colours are o f very good quality with m ost of the colours still unchanged.In BFC the paintings were executed with a very pow dery paint which, as G unn & D u Plessis (1978) noted, has changed colour in a num ber o f cases, but due to its pow dery nature, it is being sm udged and rubbed off with tim e.As a result the coarseness o f the paper is accentuated and clearly seen in the B renthurst reprodu ctio n particularly on BFC 19, Spiloxene capensis (BRI 125).The quality o f the BRI G ro u p A paintings, m ost o f which are petaloid m onocotyledons, is outstanding.The perspective in the flowers is extremely good and is far superior to any th at I have seen in the o th er early florilegia.Some o f the bulbs an d corm s have been painted in very fine detail, for exam ple, BRI 125, Spiloxene capensis; BRI 132, Homoglossum watsonium and the leaf-base in BRI 49, Urginea duthieae.But m any bulbs, corm s and bases o f plants have been done in m uch less detail.This seems, in my opinion, to indicate th a t the leaves and flowers were painted in the field with only pencil sketches o f the vegetative parts followed by a com pletion o f the colour w ork at a later stage, perhaps at cam p in the evening o r even at home.Several persons such as Petiver, W itsen, B urm an and the Breynes stated th at their paintings were executed from live plants at the Cape.The dem and for paintings in the late 1600's and early 1700's m ust have been due to the lack o f colour reproductions in books and it is certain th a t copies o f originals were m ade to satisfy this dem and.Here the enigma of the Codex W itsenii and C laudius paintings is the prime example.Copying, w hether at the Cape or back in E urope, was com m on.A glance through IPA or the reproduction o f the plates by Kennedy (1967) will show th a t the painting o f a legume occurs twice with only a few pages separating them .The folded paper o f the G roup B paintings in BFC points to the copying having been done at the Cape.It was suggested by G unn & D u Plessis

G
unn & D u Plessis (1978) state th a t the two volumes in this series o f four sim ilar florilegia, O X F and LD , both contain paintings they regard as inferiorly executed copies.They state th a t " The numbers at the base o f all the paintings . . .are in the first instance related to this volum e 'B F C ' where the num bers ap p ear chronologically and related to page num bers" .The O X F volum e is vellum -bound and has the paintings pasted o nto the page and signed " A.B. del" .They m ention th at one o f the younger Breyne's daughters had signed her own paintings " A .B ." .However, Edw ards (1964) states that this volume form s S H E R A R D MS. 188 and is a collection of paintings no d o u b t draw n by A nthonie van Breda, o f plants grow ing in gardens in H olland. Boerhaave bought the volum e for Sherard from Levinius Vincent w ho th ro u g h m arriage acquired van Breda's fam ous m useum .This relationship o f the volum e with van Breda and its very strong sim ilarity with BRI and BFC is very confusing and needs further investigation.
Fig. 4, left).In BRI, 27 paintings have com m on nam es, 23 G roup B, 4 G roup A w ith the rem ainder of G roup A w ithout a com m on nam e. BFC has the same 27 paintings with com m on nam es except th a t BFC 8, Ornithogalum thrysoides, is a quite different painting o f G roup B as opposed to BRI 52 which is one o f the 4 G roup A paintings.B FC has 28 extra paintings o f G roup B with com m on nam es.One anom ally is the occurrence o f 7 shared paintings o f G ro u p B with com m on nam es.In BRI 43, Erica cerinthoides, there is an additional set o f w ords w ritten in the same handw riting " Erica coris fo lio hispido cerinthoides africana Breynia", which is the Latin polynom ial given to the species by Jak o b Breyne in his C enturia prim a of 1678.
infralia BRI 89: Piramus in jtaliaThese anom alies w ould indicate copying from the same original, but in BFC by a rather poorer copier o f words.Jessop (1965) pointed o u t th at the com m on nam es in BRI m ust have been added to the bound volum e, as m any o f them have been blotted on the verso o f the preceding page.7Ú i t*'J -. ./ tV ^ _ ^ -^ < 3 /T 5 Jk * L \u S fic P iG 4 -Wachendorfia paniculata Burm.Top left, painting number 45 from the BRI florilegium, ><0,36; top right, Tab. 1 IX from B re y rv e s s Prodiwm (1739), x 0 ,5 .Monsonia speciosa L. Bottom left, Tab.XXI from Breynes's Prodrom.(1739), x 0 ,5 , bottom right, painting number 22 from the BRI florilegium, xO,36.
My conclusions from a com parison o f BRI and BFC are th at the BRI volum e was painted first with G roup A paintings painted from live plants and the G roup B paintings copied at the Cape from Van der Stel's collection o f paintings.They were then num ber consecutively.The same artist, or perhaps another one, m ust have copied some o f the BRI G roup A paintings for Jak o b Breyne, and yet another artist copied the G ro u p B paintings for Breyne possibly from the sam e collection o f Van der Stel.These paintings were eventually bound in 1724 and the pages num bered.The oth er two florilegia, O X F and LD m ust then have been copied from BFC.
and in the very stylized leaves o f Polygala bracteolata in the same figure.F u rth er examples o f this type can be found in BFC.