Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 1 of 9
Radiocarbon dates of dendro-dated timbers from Roman London
show large offset
Petra Ossowski Larsson* and Lars-Åke Larsson, Sweden
* Corresponding author: firstname.lastname@example.org
This article is about a rarity: radiocarbon dates of timbers archaeologically anchored
in West-Roman time which are also dated by dendrochronology. The surprising but
apparent trend is that the radiocarbon dates are a large number of years younger
than the dendro dates. This strongly supports our hypothesis that West-Roman
history and archaeology are conventionally dated too old by more than two hundred
years, and that European dendrochronology was adapted to this error already in its
Cross your heart, how do you know that the most recent two millennia of European
history lasted for two thousand years? Most probably you will answer that the written
history convincingly says it, backed by scientific evidence in the form of
dendrochronology and astronomy.
We thought so, too. Until we found that all European tree-ring chronologies
crossdating against wood samples of archaeological origin have a gap between a)
the Recent complex, reaching from today back to early medieval time around AD
400, and b) the Roman complex, archaeologically anchored in Roman time and
reaching up to about AD 200. A valid dendrochronological bridge between the well
established complexes has never been demonstrated as timber from Late Antiquity is
scarce. In order to in spite of that provide a useful dating tool for wooden Roman
artifacts it had been necessary to "calibrate" the floating Roman complex with
historical considerations. This means that the Roman dendro complex was placed
within a narrow time frame so that any archaeologically important timbers would get
historically acceptable felling dates, followed by accepting the first and best but weak
dendrochronological match against the Recent complex as the true one. Our
dendrochronological investigations suggest that this calibration made the Roman
complex too old by more than two hundred years on a real time-line (ref.1). Even
though our articles apparently are read by influential scientists, our results are still
rejected and our request for additional raw data, which is claimed to exist and which
could disprove or prove our assertions, has been ignored so far (ref.2).
Our subsequent extensive astronomical study (ref.3) showed that a scenario about
the first millennium of the Christian era being too long by a substantial number of
years was possible. Moreover, we were able to quantify this possible artificial
overstretching to be 232 years, which we suggest were inserted in the historical time-
line already when the Christian era was invented, that means in Late Antiquity in
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 2 of 9
Alexandria. The qualified and elegant manipulation was likely done by means of
astronomical retrocalculation by professional astronomers which made the fake
extremely difficult to detect.
Removing 232 years of first millennium consensus history has of course "bizarre"
consequences, for example that large parts of later West-Roman history would have
run in parallel with early East-Roman history. But before we take this scenario for
granted and dive into an alternative historical synthesis - and while waiting for
mainstream scientists to provide ultimate dendrochronological evidence - we will look
at one more scientific method which we previously have rejected as being too
unprecise: radiocarbon dating.
What about a case with radiocarbon dates of timbers archaeologically anchored in
West-Roman time which are also dated by robust dendrochronology? Now you
object that this will result in circular argument, as the time base for radiocarbon
calibration depends on dendrochronology which we claim is corrupt. But fortunately
this problem was prepared for when the first international calibration curve was
established, as we shall see below.
Radiocarbon measurement would have been the ultimate absolute dating method, if
it had worked as Willard Libby originally thought it would. I.e. all living organisms
exchange carbon with the atmosphere which has a stable 14C/12C-ratio. When the
carbon exchange stops e.g. because the organisms die, the radioactive 14C starts to
decay with a constant half-life of more than 5000 years. Thus measuring the 14C/12C-
ratio of retrieved organic matter would directly give information when the organism
died, even if this was a very long time ago.
Now the 14C/12C-ratio in the atmosphere turned out to be anything but stable. 14C is
generated in the upper atmosphere by cosmic radiation, which is highly variable.
Moreover, "old" carbon from the oceans, tundras and from volcanoes is injected into
the atmosphere at a changing rate, not to mention the burning of fossil fuels. Soon
after this unpredictable behaviour had been understood, the necessity of a calibration
procedure when converting measured "14C-ages" into true calendar ages was
realized (ref.4). For this calibration the radiocarbon content of many samples of
known ages had to be measured.
Trees suitable for radiocarbon calibration ideally form just one tree-ring containing a
lot of carbon each year. Only the tissue building the cells of the outermost ring during
each growing season is exchanging carbon with the atmosphere. Therefore a tree
forms an annually layered natural archive in which each tree-ring stores information
about the atmospherical 14C/12C-ratio of the year in which it was grown. The true
calendar age of each tree-ring can be (ideally) established with dendrochronology.
The first calibration curve was measured by Hans E. Suess from bristlecone pine
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 3 of 9
The tree-ring data sets underlying the modern atmospheric calibration curve
(IntCal13, ref.6) were in large parts published already in 1986. At that time each "high
precision" radiocarbon measurement was time consuming = expensive and required
large sample volumes. Therefore the wood of ten or even twenty adjacent tree-rings
was combined and averaged to generate just one measuring point. In this article we
are looking into details only for Roman time around AD 450 to 150 BC, and for
medieval time around AD 1500 to 1000. The backbone of IntCal13 in these time
ranges is formed by decadal (ten years combined) radiocarbon measurements from
large conifers from north-western America (ref.7). Parallel to this data runs a mostly
bidecadal (twenty years combined) calibration set measured from Irish oak (ref.8).
From AD 45 to 150 BC decadal measurements from German oak are included as
well (ref.7). The American conifer dendrochronology is regarded as "absolute" and
continuous, as is the Irish oak (Belfast) chronology. Absolute here means that a
chronology is properly anchored in recent time. We have been able to verify that the
Irish timbers used for radiocarbon calibration are correctly dendro-dated at least back
to AD 110. The German oak chronology however has a suspicious gap with very low
sample depth between AD 400 and 200 (ref.1), but fortunately only a few
measurements from this chronology are part of the early Roman time range of
Minze Stuiver, from the Seattle lab measuring the American conifer curve, concluded
already in 1982 that dendro errors might be responsible for larger offsets between
different data sets. To avoid risks he therefore rejected data derived from both
bristlecone pine and German oak for the recent two millennia of the calibration curve
The interlaboratory comparison with La Jolla and Heidelberg yields offsets
between data sets. These offsets (up to 58 radiocarbon years) are most likely
due to laboratory bias. A "real" offset, of perhaps 23 years, appears possible
for the California Sequoia and German Oak radiocarbon ages. This offset
may be due to differences in wood 14C content, but may also be due to
errors in the dendro-age determinations. [ ... ] A first attempt in averaging will
be made by combining the Seattle data with those of the Belfast laboratory
when the latter laboratory has finished the calibration of the last 2000 years
of the Irish chronology.
This makes us confident that the recent radiocarbon calibration curve largely is
based on correct dendro data over the last two millennia and thus could give us
correct radiocarbon dates of timbers archaeologically anchored in West-Roman high
imperial time (about AD 96 to 284).
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 4 of 9
Radiocarbon dates of Roman timbers
The British archaeologist Alex Bayliss wrote in 2009 (ref.10):
... [radiocarbon] studies in the Roman period remain extremely rare as there is a
perception that artifactbased dating is more precise (and less expensive!).
However, a couple of radiocarbon measurements of substantial oak timbers from
Roman sites were made in England before 1981 (ref.11). This was the time before
the fateful "calibration" of the Roman dendro complex, that means the Roman
complex was still floating in time and the archaeologists were eager to know if their
"artifactbased dating" would be compatible with scientific dating. Among the sites
excavated during that period was the Thames waterfront in the City of London. While
final dendrochronological dates for the timbers of the Roman waterfront still were
unavailable, Mike Baillie commented on the radiocarbon results (ref.12, page 255):
It is ironic that Ireland should have a completed chronology across the whole of
the Roman period and no Roman sites. In fact there are very few sites of the
first half of the first millennium in Ireland, wood-bearing or otherwise (known to
archaeologists). However, in England it is only a matter of time before one, or
several, chronologies of this period become available. These chronologies will
almost certainly be tied down by cross-dating against either the Belfast or the
Hollstein/Becker chronologies, preferably both. In England there is already a
282-year floating chronology derived from Roman waterfront timbers (Morgan
and Shofield, 1978). This chronology is of interest because on archaeoIogical
grounds it is thought to belong to the second century AD. If so, it would span
roughly the period 100 BC to 200 AD. However, there are four radiocarbon
dates associated with the chronology which in the author's opinion would be
much more at home with timbers of the very late fourth century . It will be
interesting to see if the radiocarbon dates turn out to be correct. If they do, we
will be seeing another example of the problem encountered by Hollstein.
The article by Morgan and Shofield (ref.13) actually includes a mean value curve of
the 282-year floating chronology mentioned (from New Fresh Wharf). This
chronology can be safely dendro-dated against the Hollstein chronology (but only
after more English material is added) to span -72 to 209 within the Roman complex.
This means that the radiocarbon dates turned out to be incorrect. And it was not only
Baillie who got a surprise. Harvey Sheldon and Ian Tyers wrote about another
sequence of the "City chronology" (ref.14, page 358):
This was unexpected since the Baynard's Castle piles have a series of radio-
carbon dates which have been taken to suggest felling c. ad 330-350
(uncalibrated), while the tree-ring match between Baynard's Castle and New
Fresh Wharf gives a final ring date to the Baynard's Castle sequence of AD255.
Note that calibration makes the suggested felling date even younger (about AD 395).
Now let's have a closer look at these and some other radiocarbon dates from dendro-
dated oak timbers, all from London.
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 5 of 9
14C age BP
Baynard's Castle 30-50
Baynard's Castle 55-75
Baynard's Castle 80-100
Baynard's Castle 80-100
New Fresh Wharf
New Fresh Wharf
New Fresh Wharf
New Fresh Wharf
New Fresh Wharf
New Fresh Wharf
New Fresh Wharf
Custom House 40-60
Custom House 70-90
Custom House 115-135
Trig Lane 38-48
Trig Lane 18-27
Trig Lane 8-28
Trig Lane 41-60
Trig Lane 9-28
Trig Lane 60-84
Table 1: Radiocarbon dates of oak timbers from the Thames waterfront in the City of London.
Radiocarbon measurements have been done on wood segments of a number of adjacent tree-rings
(see column Rings). All samples are also dated with dendrochronology (middle year of the adjacent
tree-rings). All information is extracted from ref.11, except a) which is from ref.14, b) which is from
ref.13, and c) which is our dating.
These radiocarbon measurements were all made at the radiocarbon laboratory of the
UK Atomic Energy Authority at Harwell (Lab ID HAR) between 1976 and 1979 and all
are uncalibrated. The radiocarbon dates Mike Baillie refers to are the first four from
New Fresh Wharf in table 1 (HAR-1864 to -68). When high-precision calibration
became available in 1986, all dates were converted, and the sample submitters were
asked to comment both on their sites and the calibrated dates.
We made however a new calibration of the measured 14C-ages using the latest
OxCal version 4.3.2 (ref.15) and IntCal13 (ref.6). The result is visualized in Figure 1.
The grey blotches show the calibrated time range in which the correct date (BC/AD)
for each sample is expected to be found. The brackets under the blotches represent
the probability for outcome within 1σ (68.2%), and 2σ (95.5 %). We also set out the
average conventional dendro age of each sample as a red peak on the BC/AD time-
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 6 of 9
Figure 1: Calibration of the measured 14C-ages in Table 1 using the latest OxCal version 4.3.2 (ref.15)
and IntCal13 (ref.6). The grey blotches show the calibrated time range in which the correct date
(BC/AD) for each sample is expected to be found. The brackets under the blotches represent the
probability for outcome within 1σ (68.2%), and 2σ (95.5 %). The average conventional dendro age of
each sample appears as a red peak on the BC/AD time-line. Note the trend within the Roman time
samples that the conventional dendro ages are 100 to 200 years older than the radiocarbon median
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 7 of 9
The graphic clearly shows the late calibrated dates for the waterfront samples of New
Fresh Wharf and Baynard's Castle suggested by Baillie, Sheldon and Tyers based on
the radiocarbon measurements. And we also clearly see that the conventional
dendro ages established afterwards tell a different story: they are all, except for HAR-
3103 and HAR-3105, between 100 and 200 years older than the radiocarbon median
values. Many of them are outside 2σ, all on the left side towards older time.
As we know the time gap between the different dendro ages of the Roman samples,
we can make a wiggle match (Tree-ring sequence in OxCal 4.3.2). The modelled end
year of the sequence has a median value of AD 373, to compare with the assumed
dendro age of AD 229. That means +144 years offset between radiocarbon dating
and dendro dating.
You could now say that this offset depends on systematic problems with these old
radiocarbon measurements, lab bias or a substantially bigger error than stated. But if
that was the case, samples from other periods measured at Harwell should be
expected to have the same systematic problems. Therefore have a look at the results
from the medieval waterfront at Trig lane at the bottom of the diagram. All the dendro
ages appear within 1σ of the radiocarbon dates! These medieval samples were
measured in the same lab, both after and before the Roman samples were
measured. They give a radiocarbon dating which is fully compatible with the dendro
dating. The modelled end year of the medieval sequence has a median value of AD
1349 at an assumed dendro age of AD 1376, i.e. -27 years offset between
radiocarbon dating and dendro dating. The large positive offset appears only with the
We do not know if this apparent trend within the Roman samples was at all noticed
when the conventional dendro dates against the Hollstein chronology became
available. What we do know however is that the radiocarbon results were taken
lightly and dismissed in favor of the dendro dates (ref.11, page 96):
The radiocarbon results were of no help in tying down the tree-ring sequence.
This was subsequently dated when more Roman tree-ring chronologies became
available. [...] The general variability of the individual 14C results suggests that
this form of dating should only he used to obtain a rough indication of the date
of a sample, not as a means of establishing an exact date.
This is the reason why radiocarbon studies in the Roman period are extremely rare,
you normally won't get any useful values but have to subtract up to 200 years from
the calibrated radiocarbon dates to get reasonable dates compatible with
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 8 of 9
Discussion and Conclusions
Instead of questioning the radiocarbon dates, it certainly would have been better to
question the dendrochronological dating of the Roman complex of the Hollstein
master. How weak the link between the Roman complex and the Recent complex of
this most important reference chronology actually is has been demonstrated recently
(ref.2). We regard the offset between radiocarbon dates and dendro dates for the
Roman oak samples in figure 1 as a direct proof that the West-Roman complex is
misdated by an embarrassingly large number of years, both
archaeologically/historically and dendrochronologically. This is not only valid for the
London City chronology, but especially for the Hollstein master and all other
European tree-ring chronologies which crossdate towards this master, e.g. the
Hohenheimer Jahrringkalender. Also the Belfast chronology has the same problem in
its part roughly covering the first millennium BC, as this part crossdates towards the
English Roman complex (ref.1).
This also means that the BC part of the modern atmospheric radiocarbon calibration
curve most probably is built on corrupt dendro data, as its radiocarbon
measurements are largely derived from Irish and German oak. To say it once more
and explicitely: Only the recent two millennia (roughly the AD-period) of IntCal13
were measured from absolutely dated tree-ring chronologies, American conifers back
to 145 BC and Irish oak back to AD 110. For older times, dendrochronologically
floating chronologies from Irish, English and German oak were used. We claim that
the conventional end dates of these floating chronologies are incorrect.
However, the use of absolute chronologies in the AD-period ensures that
archaeological samples from the West-Roman high imperial period (about AD 96 to
284) give correct calibrated radiocarbon dates. And these radiocarbon dates turn out,
at least for the London City chronology, to be 100 to 200 years younger than the
corresponding conventional dendro dates.
To resolve these problems we suggest that
a) new radiocarbon measurements of dendro dated wood samples from West-
Roman high imperial time (there are a lot of them) are made to confirm the
trend which is apparent in the Harwell data from the London City chronology,
b) a significant dendrochronlogical bridge over the Roman gap is finally
c) a project to rid the radiocarbon calibration curve of insecure dendro data is
launched, preferably before the next update of IntCal.
Meanwhile, we will continue our alternative historical synthesis with confidence.
1. Ossowski Larsson P. & Larsson L.Å. (2015). Dendrochronological dating of Roman time.
ResearchGate DOI: 10.13140/RG.2.1.5129.6806.
Radiocarbon dating of dendro-dated timbers from Roman London, draft, 2019-06-28, Page 9 of 9
2. Ossowski Larsson P. & Larsson L.Å. (2019). Response to Rzepecki et al., "Missing link in Late
Antiquity? A critical examination of Hollstein’s Central European Oak Chronology". ResearchGate
3. Ossowski Larsson P. & Larsson L.Å. (2016). Astronomical dating of Roman time. ResearchGate
4. Stuiver M. and Suess H.E. (1966). On the relationship between radiocarbon dates and
true sample ages. Radiocarbon, 8, 534-540.
5. Suess H.E. (1978). La Jolla measurements of radiocarbon in tree-ring dated wood. Radiocarbon, 20
(1), 1-18. https://journals.uair.arizona.edu/index.php/radiocarbon/article/view/522/527
6. Reimer, P. J., Bard, E., Bayliss, A., Beck, J. W., Blackwell, P. G., Bronk Ramsey, C., … van der
Plicht, J. (2013). IntCal13 and Marine13 radiocarbon age calibration curves 0-50,000 years cal BP.
Radiocarbon, 55(4), 1869–1887. http://doi.org/10.2458/azu_js_rc.55.16947
7. Stuiver M. and Becker B. (1986). High-precision decadal calibration of the radiocarbon time scale,
AD 1950-2500 BC. Radiocarbon, 28( 2B), 863-910.
8. Pearson G.W., Pilcher J.R., Baillie M.L., Corbett D.M., Qua F. (1986). High-precision 14C
measurement of Irish oaks to show the natural 14C variations from AD 1840-5210 BC. Radiocarbon
28(2B), 911-934. https://journals.uair.arizona.edu/index.php/radiocarbon/article/view/1004/1009
9. Stuiver, M. (1982). A high-Precision calibration of the AD radiocarbon time scale: Radiocarbon,
24(1), 1-26. https://journals.uair.arizona.edu/index.php/radiocarbon/article/view/740/745
10. Bayliss, A. (2009). Rolling Out Revolution: Using Radiocarbon Dating in Archaeology.
Radiocarbon, 51(1), 123-147. doi:10.1017/S0033822200033750
11. Jordan D., Haddon-Reece D. and Bayliss A. (1994). Radiocarbon dates from samples funded by
English Heritage and dated before 1981. English Heritage, London.
12. Baillie, M. G. L. (1982). Tree-Ring Dating and Archaeology. (Croom-Helm, London)
13. Morgan R. and Schofield J. (1978). Tree-rings and the archaeology of the Thames waterfront in
the City of London. In Dendrochronology in Europe, Fletcher (ed.), BAR, International series, 51, 223 -
14. Sheldon H. and Tyers I. (1983). Recent dendrochronological work in Southwark and its
implications. London Archaeologist 4:13, 355-361.
15. OxCal online version 4.3.2, Christopher Bronk Ramsey 2019. https://c14.arch.ox.ac.uk/oxcal.html