Retrieving some Hollstein data
by Lars-Åke Larsson

The Hollstein Middle European Oak chronology

ISBN 3-8053-00964
In 1980 Ernst Hollstein published his book "Mitteleuropäische Eichenchronologie" (Middle European Oak chronology).
It is a 273 pages book with pages in the size of 315 * 225 millimetres where the author presents a chronology going from 724 BC to 1974 AD.

Today the chronology is considered outdated mainly because of some errors and deficiencies in the part before AD 1000.

For an amateur without access to kept-secret university tree-ring data, the Hollstein data may anyhow be of interest, especially as it contains not only a chronology data set of one data per year, but also a set of overlapping ring width curves which can be converted to a collection of individual samples. Such a collection gives us standard deviations for the chronology which makes crossdating easier and more reliable. The drawback with these curves are that many of them represent mean values from several individual samples. Consequently we have no chance to correct any errors done when the samples were laid together to calculate the mean value series.

Another aspect on this data and on this book is that it gives an interesting survey of historical wood found in various constructions. There are detailed information on many such wood samples and their background.

The mean value chronology

There are eight columns with regional mean value chronologies and also a chronology for all of them together. The regional chronologies cover various periods of time. I have only transcribed the rightmost column with the total mean value chronology.

Each value is given as a "Wuchswert", i.e. as a "growth value".
These values are calculated from ring width values w(1), w(2), w(3) etc as Wuchswert(n) = eLog(w(n)/w(n+1))
This algorithm for normalization of data can be selected within CDendro as the Hollstein normalization.
Though to make the values from the book fully compatible with CDendro, they have to be scaled.

The printed quality is not very suitable for any automatic read-in without extensive proof-reading. The data can be converted to a format readable in CDendro, e.g. as:
 
-0.003 #1658 
0.007 #1657 
-0.178 #1656 
0.173 #1655 
0.137 #1654 
0.058 #1653 
-0.165 #1652 
-0.136 #1651 

For your research purposes, that file is made available here: HollsteinWuchswertWE.d01
You can open it in CDendro as Samples/Open normalized file, if you first change your settings to use the Hollstein transformation.
N.B. Note that Hollstein zero values for "Wuchswert" should be converted to 0.001 to make CDendro not handle it as an inserted zero-ring.

Note: Please note that the content of this data file is my interpretation of the data from the book - By mistake, I may have introduced some errors!

Let us look for repeating ring width patterns!
See that Enable crossdating quality test functions is checked, and also see that there is no other sample selected as the reference!
Click the Calc new diagram button shown above! - When asked, specify 100 as block length!
100 years long blocks (segments) of the curve have been compared to all possible positions along the curve.
The best matches (which are not the right ones) are plotted in the curves above. The red curve shows correlation coefficients, the blue curve shows T-values.

There are some peaks where T-values go as high as 6.8!
This means that the Hollstein chronology curve has some repeating patterns! This is not normal!
What may have happened is that one or more samples of late times have been found together with older material and then summed together as all being that old.
This would have copied one or more late-time "ring width signatures" into old time of the chronology.

The ring width curves

The red points show CooRecorder measurement points
The book contains 25 pages with plotted curves showing ring widths from samples used or mean value ring widths from groups of samples used. Normally a curve spans over more than one page.

The (vertical) ring width scale is logarithmic, but only a few curves are directly connected to a scale. So ring widths for curves without a connected scale can only be reconstructed as relative ring widths, i.e. as scaled with an unknown factor. As few curves have a connected scale, I decided to make all reconstructed ring width curves relative with all having their smallest ring width set at 0.35 millimetres.

I have measured all the plotted curves with CooRecorder from scanned pages from the book. The CooRecorder coordinate files have then been processed (with a special program, HollsteinCurveToWid) to create a relative ring width file for each curve usually spanning more than one page. The result is a ring width file (a CDendro .wid-file) for each sample.

There were some 27000 points to measure! To make measuring faster, I wrote a pattern recognition mechanism for CooRecorder which tracks curve segments from left to right automatically, though every point had to be checked visually for correctness and misplaced points replaced by hand.

Missing points in curve drawings There are a number of cases where there are not nine small rings on a curve segment between each vertical line as shown in the diagram above. In these cases, I have considered this as a drawing mistake and corrected it by inserting the missing point in the middle of the straight line.
Examples of this is: Steinbach (year 752), Mainz Brand (665), Berlegem (Beerlegem) (537, 548), Forchtenberg (472), Wederath (243), Trier-West (125), Hallstatt (-718, -709, -700, -679).

Note that year 0 does not exist, though it exists within CDendro, to avoid complicated mathematics. This makes Hollstein datings differ from CDendro for times BC.

Hollstein states that some of the curves used by him are not plotted in his book because of space limitations. E.g. in the case of Corvey (1082) only 32 rings are plotted though page 55 of the book states there should be 81 rings available. So somewhere there may be more material available.

After the .wid-files were created, I copied them all into their right places (according to Hollstein) in a CDendro collection. This was written out as a Tucson/Decadal .rwl file. A copy of that file is available below:

Hollstein relative ring width data retrieved from drawn curves:
For your research purposes that file is made available here: HollsteinWID.rwl

Note: This data is my interpretation of the Hollstein curves and their relative positions. I may have made mistakes! So when you analyze this data and find something unexpected, then do check the book. Do notice that the current new version of CDendro can plot curves the same way as Hollstein did, i.e. with old times to the left. (See Settings/Plotting window settings in CDendro.) That feature is great if you want to check and compare the curve shapes of the Hollstein book with the curve shapes displayed in CDendro.

Code identities used for the Hollstein curves

CodeName and comment
AADDAachen, Domdach, 1350-1284. MATCH AT 1768!!
AAGHAachen, Granusturm und Rathaus 1330-1150
AAGRAachen, Granusturm, 777-601
AAHLAachen, Haus Löwenstein, 1656-1461
AAHOAachen-Hof, 63- -123
ABBMAltenberg bei Müsen, 1224-1099
ALT1Altrip, 691-494
ALT2Altrip, UM 369, 350-305, ????
ANBL(Antwerpen), Bildtafel Liebesgarten, 1550-1422. MATCH AT 1487!
With only the best crossdating members kept in the reference, this sample
matches best to 1487 (corrcoff=0.39). 1550 is then NOT to be considered (corrcoff=0.19)!
The Rubens workshop may as well have reused old wood, anyhow splint is not available!
Best match towards sum of published Neth+Belgium curves say 1487 at 0.37, 1550 at only 0.26.
ARGWArgonner Wald, 1965-1818
BDN1Bad Nauheim, Siebel (Karoling), 792-653, ?? Though could be dated to 792 with corr=0.40 if extra rings are inserted at year 730 and 686.
Otherwise the corr coeff to the rest of the collection is only 0.12!
BDN2Bad Nauheim, -64- -502
BEFOBefort, -429- -523
BENGBengel, 342-274, ???
BERLBerlegem, 550-284 (??)
BICKBickenbach, 152-76, ???
BLEIBleialf, Hof, 1750-1650
BRARBraunschweig, Altes Rathaus, 1511-1450
BRBRBroye Brücke (see LATE)- Nordschweiz, 129- -23
BRENBrenig 1550-1450 -Bad match at old end.
BROIBroichweiden, 365-129. ?? 264?
BRP1Braunschweig, 1274-1094
BUNOBurg Nordenau, 1088-959
BURABüraberg 1144-1050
BUREBüdingen, St Remigius, Schloss 1113-816
CHRIChristenberg, -419- -483
CORVCorvey, 1082-1050. TOO SHORT!
DHRODhron Röm. Strassenbrücke, 234-119
DIHEDittelsheim, Kirchturm, 1053-904, ?? Matches towards 1053, 1029 and 842 of Hollstein
Matches towards year 831, 1087 and 1029 of belg003.crn
DRUGDrüggelte, Kapelle, 1150-963
DUSADüren (Saar), Humburg, 907-746
ELLAEllar, Hof, 1650-1550
ERP1Erpfingen I, 492-400
ERP2Erpfingen II, Siedlung, 643-410
ETZBEtzean, Brunnen, 1339-1269
FB16Forchtenberg Brunnen 616, 616-573, TOO SHORT!  matches FB27 (same site?) with corr=0.56 as dated.
FB27Forchtenberg, Brunnen um 627, 610-410
FILKFilsch, Kirche, 1780-1615
FRECFreckenhorst, Baumsärge, 966-788
GHKPGelnhausen, Kaiserpfalz, 1182-1063
GKZBGrosskrotzenburg, 134- -13
GRIDGriesheim bei Darmstadt, 691-550
GUSTGustavsburg Mainbrücke, 163-76, ??? Also match at Hollstein year -365.
HALLHallstatt, Salzbergwerk, -655- -721 ??  5 missing ring symbols in curve...!!!
HERBHerborn Dillbrücke, 1150-979
HETRHeidenburg, Trier, Saar 1974-1629
HUFIHüfingen, Grabkammer, 606-450
IRRBIrrel, Röm. Brunnen, 78- -123
K9SANeunkirchen (Saar), Röm. Wasserleitung, 211-41
KAWAKirche Arnoldsweiler, bist. Aachen, 967-788
KBMBKoblenz Moselbrücke, 203- -109
KDCGKölner Dom Chorgestühl, 1311-1120
KDGBKöln, Domgrabung, 314-232
KIBGKirchberg, Kirche (Hunsrück), 1750-1503
KIRNKirn-Sulzbach, -442- -690, ???
KKNAKöln, Knabengrab um 536, 511-273, ???
KMBQKöln, Mühlenbach, Quaderbau 4- -117, also match at -53!!!???
KOLEKöln, Rhein (Eisbrecher?), 1631-1482
KORBKöln Rheinbrücke, 336-149
KORHKöln, Rhein 61- -23
KORNKornelimünster, 1314-1173
KSBAKirche Stockheim bist. Aachen, 950-738
LAGELage (Lippe), -166- -223, ???
LATELa Tène+Thielle Nordschweiz, -23- -391
LESSLessenich ü. Bonn, Kirchenfenster, 1069-902
LKBBLimeskastell Butzbach, Brunnen im Lagerdorf, 183- -23
LKSBLimeskastell Saalburg, 145- -54
LKSHLimeskastell Schierenhof, 141- -70
MANCManching, -123- -348
MB16Mainz, Brand, 716-582 One ring lost in Jahrringtafeln drawing at year 665.
MB253Mainz, Brand, 253-76
MB259Mainz, Brand Röm. Brunnen, 259-176
MB53Mainz, Brand 653 n.Chr., 653-439
MB73Mainz, Brand 673 n.Chr., 673-481
MBRUMainz, Brücke, 76- -23
MBUFMainz, Brand (Rheinufer) 31- -92
MEERMotte Meererbusch,1050-904
MESCMeschede 1967-1851, drei Eichen
MEWBMeschede, St Walburga, 896-738
MGBNMönchengladbach, Neusserstrasse, 1606-1500
MGNWMönchengladbach-Neuwerk, 1150-1050
MIN1Minden, Dom, Fenster, 913-793, 1029???  Jan 2016: 1029 now confirmed! See Errata and news related to the Hollstein data
MIN2Minden Domfenster, 1062-970
MIWAMittelstrimmig Röm. Wasserleitung, 228-100
MOHUMotte Husterknupp, 964-837
MRBRMainz Röm. Rheinbrücke, 49- -88
MUEBMünstereifel, Basilika, 1099-822
MUESMünstereifel, Baumsarg, 682-521
MUROMünstereifel, Roman. Haus, 1167-1072
NUBKNürburg, Burgkapelle, 1499-1385
OB2BOberflacht, 2 Baumsärge, 550-401
OB6BOberflacht, 6 Baumsärge, 574-398
OBEROberscheld, 1531-1410
OFFBOffenbach (Dillkreis), 1358-1250
OWSMOberwesel, St. Martin, 1435-1350
PABUPaderborn Busdorfkirche, 1250-1054
PALZPalzem/Stadtbredimus, 34- -209
PBBFPetersberg bei Fulda, 832-639
PUTMPüttlingen, Mühle, 1538-1360
REMBRemblinghausen, Hof, 1704-1569
SAARSAAR 1950-1792
SCHASchalkenmehren 1442-1301
SCHFSchweich, Fährturm 1791-1690
SCHWSchankweiler, Kirche, 1732-1650
SGBBSiegburg, Brücke, 1512-1372
SINDSindelfingen, 1425-1249
SINZSinzig, Ahrbrücke, 1217-1097
SK23Stuttgart, Königsstrasse 23, 1808-1634
SLBUSchloss Büdingen, Pallas, 1668-1350
SOLLMarkoldendorf im Solling 1950-1808
SPEDSpeyer Domfenster, 1450-1350
SQUSSeine-Quelle, Statuen, 33- -123
STEDSteinbach, Einhartsbasilika, Dachstuhl, 1150-1050
STEISteinbach im Odenwald, Einhards - Basilika, 814-474
STSBStrassburg, Neuer Markt, Kanal 995, Rep. 1049, 1029-889 ????!!????
STTSStrasbourg, Tour St. Catherine, 1347-1251
TATETrier Amphiteater (Eichen), 286-70
TATTTrier Amphiteater (Tannen), 294-210 ???
THIEThielle, -223- -545
TITETitelberg (Luxemburg), -140- -223
TMB11Trier, Moselbrücke I, 71- -23
TMB12Trier Moselbrücke I, 245-76
TMB21Trier Moselbrücke II Pfeiler 5, 144-76
TMB22Trier Moselbrücke II Westturm, 157-76
TMB3Moselbrücke, Trier, 1807-1550
TMK1Trier, Moselkran, 1762-1565
TMK2Trier, Moselkran, 1850-1770
TPNITrier, Porta Nigra (St. Simeon), 1038-814
TRAMTrier Amphitheater Arenakeller, 668-442
TRD1Trier, Dom, 1450-1124
TRD2Trier, Dom, 1749-1450
TRDSTrier, Dom, Schatztruhe pg 163, 1045-728 ?? Oldest years should perhaps be cut away!
TRDWTrier, Dom Westbau, 1050-850
TRGMTrier, Gasthaus Mustor, 1839-1750
TRHBTrier, Haus Britanien, 1335-1150
TRHMTrier, Haus Marmagen 1283-1150
TRK1Trier Kutzbachstrasse 1, 1323-1250
TRM2Trier, HS. Marmagen II, 1350-1312, TOO SHORT!
TRM3Trier, Haus Marmagen III, 1647-1350
TRM4Trier, Haus Marmagen IV, 1750-1650
TRM6Trier, Haus Marmagen VI, 1825-1750
TRMBTrier, Moselbrücke 1347-1150
TRMB31Trier Moselbrücke (Römerbrücke) 285-246, SHORT!
TRMB32Trier Moselbrücke (Römerbrücke), 315-183
TRMUTrier, St. Matthias, Uhrpodest, 1599-1461
TRPBTrier Petrisberg, -29- -96
TRRBTrier Römerbrücke, -23- -339
TRROTrier Römersprudel, -349- -451
TRRSTrier Römersprudel, 141- -123
TRSCTrier, Schiffsmühle, 1287-1220
TRSMTrier, St Matthias, 1115-999
TRSOTrier, St. Matthias, Orgeldoxal, 1242-1150
TRSSTrier, Schiffsmühlensteg, 1450-1274
TRSTTrier, Sternstrasse, 1475-1351
TWMLTrier-West Moselufer, 231-32
UNLUUnterlübbe, 66- -154, ???
WACHWachtendonk, 1252-1164
WEDEWederath Röm. Brunnen, 272-66
VIDTVillingen, Tannenspaten, -503- -572, ???
WIEDWiedenbrück, Baumsärge, 916-824
VIG1Villingen, Grab 1, -550- -637
VIG6Villingen, Grab 6, -524- -650
VIG7Villingen Grab 72 u. 74, -535- -723
VIHUVillingen Hügel, -588- -639
WINKWinkel (Rhein), Graues Haus, 1078-877
VINPVillingen, Nördl. Pfostenreihe, -549- -659
VISPVillingen, Südliche Pfostenreihe, -545- -581
VITAVillingen, Tafelbild Abendmahl, 1589-1450
VIUN1No name, -544- -613, ???
VIUN2No name, -544- -569, ???
VOLKVölklingen, Wasserleitung, 1823-1698
XANEXANTEN 1967-1851, drei Eichen
XANTXantener Dom, Chorgestühl, 1225-993
ZOBBZöbingen Baumsärge, 559-383

Continue with the next section!

August 22 2008. Lars-Åke Larsson (Updated 21 Febr 2009 and 19 Jan 2016)


Copyright © 2016, Cybis Elektronik & Data AB, www.cybis.se