It is almost 25 years since I first arrived in Orkney on a foggy
day in May 1979 to gain my first experience of archaeology as
a foot soldier in the trenches of Howe. I'm not sure now of the
first of the many occasions during that quarter century when I
first met Daphne, but they have been invariably enjoyable because
of her contagious enthusiasm.
If only there was a way of capturing this enthusiasm in bottles
which could then be uncorked in dreary meetings to inspire all
those present with her fervour for archaeology
Here, for Daphne, are some thoughts about the significance of
Howe two and a bit decades later for brochaholics and others interested
in the prehistory of Northern Scotland.
Thanks are due to Anne Brundle and Beverley Ballin Smith for
encouragement and comments, to Martin Carruthers for some suggestions
which made me go back to some reports I'd assumed I'd read completely,
and to Charles Tait for providing the illustrations.
Part 1: Howe
Howe was excavated from 1978-82 as a rescue dig ahead of agricultural
improvement. The mound in the field at Howe Farm near Stromness
had been recognised as a probable broch site from its visible
size, but other than records of a Norse glass linen smoother found
in the 1860s and some other Iron Age finds from antiquity (Ballin
Smith 1994, p259), there was no indication before excavation of
what else the site might contain. This is a point worth highlighting;
the Howe mound was typical of the many such lumps in the Orkney
landscape and elsewhere, with no real indication of the complex
sequences it contained.
The excavation revealed use of the site, originally a post-glacial
boulder clay knoll, over perhaps 5,000 years until around the
Norse arrival. Crucially, although the broch period phases provided
the most massive structural features of the chronology, Howe was
not a broch site as such, but a site which contained broch constructions
within a much greater chronological sequence. Howe was excavated
down to its Neolithic layers; it is still the only occasion when
a broch has been completely removed in a controlled excavation
with modern standards of information recovery. The potential insight
is therefore unparalleled; the results of that insight into the
broch period are also unique when compared to other broch investigations.
These thoughts are designed to highlight whether Howe was a unique
aberration, or whether it can, because of the comprehensive nature
of its investigations, provide new lines of enquiry for broch
Potted summary of discoveries
The excavation report (Ballin Smith, 1994) has been selectively
filleted for the salient points here, including the identified
phasing descriptors. Howe had two main Neolithic phases (Phases
1 & 2), culminating in a massive remodelling of the site to
create a chambered cairn with surrounding ditch and bank. The
earliest post Neolithic use of the site is unclear (Phases 3 &
4) due to its extremely fragmented survival; the remnant evidence
consisted of gullies truncated by later features, two wells and
floor surfaces from perhaps three huts, as well as significant
stone-robbing of the chambered cairn. However their importance
is in proving that the site was in use during that period, and
that whatever the use and vanished structures involved, the central
feature of the pre-broch site was sufficiently important to warrant
significant modification and maintenance. Phase 5 saw the construction
in the Early Iron Age of a roundhouse on the remnant clay core,
along with some surrounding buildings; this entailed the obliteration
of much of the earlier features, particularly of the Phases 3
& 4 structures. Whether the roundhouse was simple or complex
in form is unclear, but it was probably sufficiently massive to
require extensive site clearance and preparation before and during
Despite this site preparation, the structure showed signs of
collapse and instability, and was substantially remodelled in
Phase 6 to create the first definite broch. This was built to
a (so far) unique combination of architectural features and was
in many ways an exquisite and elegant piece of design engineering
which has been overlooked in broch commentaries since. The broch
was accompanied by a complex of external settlement structures
within the area enclosed by the remnant Neolithic ditch.
The instability issue resurfaced with further collapse evident,
and Phase 7 saw another substantial reconstruction to create a
second more massive and architecturally more conventional broch,
again with external buildings forming a typical Orkney broch village.
This remained in use for a long period into the Late Iron Age
and beyond, despite further collapses of the broch and its abandonment
as a functional building.
|Figure 1. The unexcavated broch mound at
Simbister, c. 3 miles SE of Thurso, about which nothing is
known (photo by David Lynn).
Three main themes emerge from this summary. The first is that
while an unexcavated broch mound can be diagnosed as having a
broch or similar feature at its core with a high degree of accuracy,
the other components of the site's sequence will be largely hidden
until excavation (Figure 1).
As a side comment, this is also true of many broch sites which
are exposed by other means such as marine erosion; while non-broch
components may be visible - particularly for post-broch sequences
on the site - their accurate diagnosis is impossible without precise
exploration. I want to focus here on the potential for pre-broch
sequences, so will add that excavated and preserved sites give
very little clue to their pre-broch state if they are displayed
at the broch levels; this point will recur later, but is a strong
perceptual distortion due to the prominence of guardianship sites
which have (justifiably) been explored only down to the level
of the most massive structure in the sequence, eg Gurness (see
|Figure 2. Aerial view of Gurness, showing
how a monument conserved for display can appear to be a single
phase complex and conceal any evidence for its preceding sequences
(photo by Charles Tait).
The second is to recognise the significance of pre-construction
Howe showed the obliteration of much of the immediately preceding
structural remains, with only isolated and subtle clues to the
presence of an earlier sequence which was only identified because
of the complete removal of the broch structures. In a general
sense, this is a logical scenario; a broch is a massive construction
which would require prior preparation of the site to allow a satisfactory
build. It is also a logistical issue, as importing the huge quantities
of construction material (particularly the building stone) to
the site and arranging its deployment on site would be impractical
unless space was cleared for access, manoeuvring and preparatory
shaping. The irony at Howe is that while pre-constructional clearance
was massive, the decision to retain the remnant clay core of the
Neolithic cairn - rather than remove it - caused a perpetual instability
blight which was never resolved despite its damaging consequences.
|Figure 3. Howe of Howe during excavation
in 1981. The Phase 7 broch is being removed, showing how the
Phase 6 broch emerged from within the ringwall. The picture
also represents the chaos accompanying a major remodelling
of the site; while a controlled demolition is underway through
the excavation process, it is perhaps the closest illustration
of what a constructional remodelling might have looked like
with its accompanying disruption of extant features (photo
by Charles Tait).
The third theme is that to all intents and purposes the second
broch (Phase 7) appeared to be a "normal" example as
it was gradually revealed by excavation, with no significant clue
to the sequences it had undergone, particularly in Phases 5 &
6 (Figure 3).
If excavation had stopped at that point to create a monument
for display, none of these would ever have been realised. Many
preserved and displayed brochs now exist, looking as if they were
apparently a coherent single construction. It has to be a real
possibility that in many of these cases this is an illusion, not
a reliable indicator of their true form. Again this possibility
has a general logic to it if the duration of the broch period
is considered. Many broch sites apparently show a relatively intact
central broch until abandonment at the end of the Middle Iron
Age; this relationship is mainly derived from the presence of
Roman artefacts in later layers due to the absence of C14 and
other absolute dating evidence for the vast number of sites explored
between 1850 and 1950. If the first broch-like structures (ie
massive-walled round buildings, albeit without the complex architectures
of the now surviving towers) first emerged c 700BC (eg Bu of Navershaw,
Quanterness, Pierowall), it is unrealistic to expect that these
structures could survive up to 8 centuries of use and wear without
major reconstructions during their individual chronologies.
Part 2: a wider context
Orkney's missing Bronze Age?
The next ingredient of this account is to look further back into
Orcadian prehistory, and at a paradox in our understanding of
the Late Bronze Age. All the evidence suggests that Orkney was
well settled in the Neolithic, with a wealth of surviving remains
corresponding to a period of climatic optimum, and continuing
into the Early Bronze Age. At some later point, perhaps around
1200BC the climate deteriorated dramatically for 500 years or
so, possibly related to the violent eruption of Hekla in 1159BC.
This saw the collapse of the Neolithic land patterns, and apparently
left us only with the burnt mounds phenomenon as evidence for
what was happening in that missing five centuries before the Early
Iron Age. Although over 230 burnt mounds are known in Orkney,
situated in or near "agricultural land" (Hedges 1974,
pp 61-65), they are not easily explained, particularly in a search
for the main form of house type for the period. Instead there
is an impression that we are missing a key structural form for
these few centuries, a type of building which would have been
more clearly residential in design and which would have occurred
in sufficient numbers to be recognised today in normal archaeological
The paradox is that for a period of climate collapse, which was
colder and wetter than any post-Ice Age era, the folk in Orkney
apparently chose to live in the boggiest and marshiest parts of
the landscape (exaggerating to some extent from Hedges 1974, p79)
which is where we find much of the Orcadian burnt mound distribution
today. Something feels wrong with this explanation, but we have
no archaeological evidence to correct it as things stand. There
are no other recognised remains from the Orcadian Bronze Age in
the quantity needed to provide an alternative.
The insight here is to suggest that they were there, but no longer
exist in a recognisable form. One clear possibility is that they
were in fact on sites on which brochs were later built. If so,
their chances of survival would be very low in the site reorganisation
prior to broch construction, and the chances of discovery today
would be even lower as it would entail controlled removal of the
broch to reveal what was left immediately underneath.
Although a hypothesis which by its nature is largely untestable,
this suggestion does give a scenario to fit the Howe Phase 3 &
4 sequences - scattered incoherent strands of evidence across
the site's extent without a definitely identifiable form. Again
it is worth emphasising that Howe is the only excavation so far
with a format and scale to enable such evidence to be uncovered
Other relevant broch sites
Having said that the Howe strategy was a unique demonstration
opportunity, it is also worth examining other sites to see if
any comparable evidence can be identified.
Perhaps the closest comparison is the multi-period site at St
Boniface on Papa Westray, where a vertical eroding cliff section
was investigated and recorded. An Iron Age roundhouse again sat
within a sequence of preceding and subsequent structural deposits,
with some remains which were dated to the Late Bronze Age (Phase
4 of the site's chronology) before the broch-like roundhouse (Phase
5). The excavator's difficulty in establishing the exact stratigraphic
relationships between features in these two phases (Lowe 1998,
pp 117-8) may well be compatible with a Howe-like obliteration
of immediately pre-roundhouse features.
Tofts Ness on Sanday has yet to be published, but preliminary
findings were of a massive Early Iron Age roundhouse on the site
of a removed earlier roundhouse, which was almost certainly a
shaggy roof design (ie the eaves of the conical roof extended
beyond the outer wall as was common in much of Britain); the evidence
was provided by drip channels for water draining from the eaves
in the ground surface at the foot of the outer face of the ringwall,
and a C14 date of c 1000BC came from associated deposits (Steve
Dockrill, pers comm.).
Bu of Navershaw, a frantic salvage excavation of a massively
walled broch-like structure, showed some relevant features. The
first phase of the site, Phase 1a, was an "occupation of
unknown extent and nature" (Hedges 1987, v1, pp5-10), lying
outside the extent of the later roundhouse. It was sealed by a
ploughsoil horizon in Phase 1b, which underlay the roundhouse,
implying a significant time gap or a major remodelling phase.
However the floor of an earth house subsequently inserted into
the roundhouse some time after its abandonment produced animal
bone with a C14 date some 100 years before the broch construction
(ibid, p93) suggesting that the earth house insertion penetrated
stratigraphy from the immediate pre-roundhouse period. However
the rushed nature of the excavation meant no other exploration
was made of what might have been under the roundhouse.
Crosskirk in Caithness showed elements of pre-broch site use
during excavation, both outside the broch as a possible promontory
fort and - crucially for this paper - underneath the broch (Fairhurst
1984, pp56-7), though these were not fully explored by controlled
removal of the broch before the site was destroyed. There are
indications from other Caithness broch sites (albeit unexcavated)
that many brochs may be sitting on a depth of preceding rubble,
though with no clue to the period(s) of earlier origin (Swanson
Other broch sites have shown pre-broch standing structures, usually
in the form of some sort of gatehouse. Clickhimin is the most
widely recognised example (albeit from outside Orkney), but two
potential Orkney candidates are Borwick (Lamb 1980, p77) and Midhowe
(ibid, p80). Both these are unexcavated, and it is unclear in
their current unexcavated states whether they are solid or hollow
(ie ramparts or blockhouses), let alone their chronological relationships
with the broch on each site. The potential for each of these sites
to have had pre-broch settlement evidence under the broch is clear,
with the continuing proviso that the broch insertion could have
destroyed much of the evidence.
Gurness has some unexplained stratigraphical issues, suggesting
the presence of features which predate the currently displayed
broch complex, particularly for structures encroaching into the
inner ditches and ramparts. These relationships suggest the inner
ditches may predate the displayed broch complex and belong to
an earlier phase of site use (Beverley Ballin Smith, pers comm.),
which could still be Early Iron Age or thereabouts, as ditches
and ramparts are not known from earlier Orcadian settlement types.
Shetland has two other broch sites to mention in this context;
Jarlshof, with its extensive sequence of pre- and post-broch structures
(though not excavated or currently displayed in a fashion which
would add specific results to the arguments here), and Old Scatness
where a broch under current excavation is centrally positioned
within a remnant field system (Steve Dockrill pers comm.); it
is unlikely however that the excavation strategy at this site
for eventual display will produce specific evidence of the type
Non-broch Orkney sites potentially from this
The intention here is simply to establish that a number of sites
are known in Orkney which do not have brochs, but which can, on
the present knowledge base be regarded as potential contemporaries
of the broch period. In most cases these have not been securely
dated, so their listing here is provisional. These include the
Calf of Eday roundhouse (which may be a quartet of structures
- below), a wheelhouse revealed but not excavated at Lamba Ness
on Sanday in the 1980s by Bradford University as part of the Pool
project, Pool itself (which is securely dated to this period,
but has no massive roundhouse or similar feature), Skaill (in
Deerness, with a similar picture), the Minehowe complex of structures
which is still being investigated, the two thin-walled circular
huts at Spurdagrove in Harray (Hedges 1987, v3, p36), the Little
Howe of Hoxa (South Ronaldsay), and Howmae (an amorphous series
of structures including wheelhouses on North Ronaldsay; RCAHMS
At a generic level, Lamb identifies 12 possible Iron Age promontory
forts in Orkney (1980, pp 76-80). Those at Midhowe and Borwick
are discussed above, while another of his candidates - Riggan
of Kami - is actually a ground-galleried broch. While other examples
may not ultimately remain on this list and it may represent a
morphological grouping rather than a chronologically defined phenomenon,
it is still a group of sites to bear in mind.
Similarly, Orkney's many souterrains have a potential significance
as a group of monuments, again with the proviso that they may
represent multi-period phenomena rather than a tight chronological
What should be apparent from this section is the cogent possibility
that there were non-broch/roundhouse elements to Orkney's Early
and Middle Iron Age. Some known candidates can be suggested, although
our current knowledge of them is sketchy, let alone any ability
to draw generic analyses.
Areas in and around Orkney without broch sites
If this paper's hypothesis that we do not find convincing Late
Bronze Age house structures because they were obliterated or concealed
by later broch or roundhouse insertions is pursued, areas in and
around Orkney without brochs could well be expected to produce
some features of interest.
Unfortunately, the picture is inconclusive with no obvious leads
to follow from published reports. Within Orkney, Eday stands out
as the clearest broch-free area, despite being surrounded by islands
with significant broch or roundhouse distributions, particularly
Westray, Sanday, Stronsay, Rousay and Shapinsay. There is no obvious
explanation why Eday seems to have missed out - while coastal
erosion and localised peat build-up may be relevant factors, the
odds of these destroying any recognisable traces across a potential
range of broch sites are unconvincing. Yet Eday does not seem
to have an undue proportion of other sites which suggest themselves
as candidates for Early or Middle Iron Age settlement (eg RCAHMS
1984), despite 16 burnt mounds. Even Calf of Eday has one definite
and up to 3 possible roundhouse and associated structures (ibid
p18; Hedges 1987, v3, p36).
While East Mainland and the South Isles do have brochs (particularly
Hoy and Burray in some quantity in relation to the agricultural
land) the distribution is very sparse compared to West Mainland.
The sites of Minehowe and Skaill (Deerness) have shown Iron Age
house structures, but so far without producing generic information.
The high number of "enclosed settlements" identified
in RCAHMS 1987 (pp24-6) which can not be further diagnosed may
Finally, the absence of brochs or similar structures on Fair
Isle is worth noting, along with a proportionately high number
of burnt mound sites - eleven - including possibly the UK's largest
example at Vaasetter (Hedges 1974, p61). Again, the significance
of this observation is unclear.
So on current information, the idea that areas of Orkney with
a low or nil distribution of broch sites could produce clear signs
of an alternative form of Iron Age settlement presence can not
be productively pursued.
Part 3: taking these ideas forward
Brochs: greenfield or brownfield sites?
My original thoughts were that this article would culminate in
this question for consideration, and by and large this remains
the case, though some other threads can be drawn (below). The
starting point was two-pronged, firstly that Howe showed some
very specific results which deserved amplification in a wider
context, and secondly a jolt of self-realisation about complacent
assumptions of homogeneous construction gained from several visits
to displayed broch sites - or to put it another way, a visit from
the What-If? Pixie.
So, what if brochs were a phase of use of a site rather than
a determinant? The indications in an Orkney context are convincing
that many sites were previously occupied as settlements, probably
in some cases until the point at which a broch was constructed.
This does not in itself mean that a given site was continuously
occupied from the Bronze Age or earlier into the broch period.
Strict continuity of settlement is a very elusive strand of evidence
to pursue within this sort of context, when massive site remodelling
phases destroy conventional stratigraphic relationships and obliterate
much of the absolute dating materials. This is particularly significant
for arguments such as those recently made by Cowley that no evidence
was produced at Howe to show continuity (in Downes & Ritchie
2003, pp 79-80); while a strict reading of the evidence from Howe
allows Cowley's re-interpretation, it is unclear how a realistic
and watertight dating strategy could be constructed to cope with
site reorganisations of this scale to show that there was never
a period of, say, 20-25 years (ie one generation) when the site
was not occupied within the 1000 year lifespan of the Late Bronze
Age to Late Iron Age. Instead common sense has to prevail with
the acceptance that while absolute continuity can never be proven,
we should not lurch to the other end of the scale and assume occupation
was chronologically sporadic, an avenue which unleashes all sorts
of implausible arguments about roving population groups.
Cowley also underplays a crucial factor compared with sites elsewhere
in Scotland, which is the comparable durability of building materials
within the structures; all structural parts of a crannog for instance
are wooden, so will deteriorate at a consistent rate with each
other. Therefore the whole building has a finite life and has
to be replaced every few decades. In contrast, a stone building
with wooden roofing and internal fixtures has a basic shell which
can remain viable for several centuries, through successive episodes
of replacement of the wooden components, each of which would involve
probable disruption to the preceding internal deposits without
indicating abandonment or intermittent use. However excavators
of monumental stone-built structures with extensive sequences
will need to recognise the generic questions Cowley raises and
develop their strategies for information recovery accordingly.
So the question of Orkney broch constructions as greenfield or
brownfield projects has to be considered without the distraction
of strictly continuous site usage, but in a more general sense
of ongoing or recently preceding site use compared to virgin development.
Earlier parts of this paper summarised as much evidence as could
be found for broch construction on sites with evidence for previous
use, and my belief is that this was a common scenario in Orkney.
Certainly the evidence for previous use would have been readily
apparent for broch-building groups, and does not seem to have
been a deterrent.
One subtlety which is worth exploring is that those surviving
initial broch sites in Orkney - which may well be better expressed
in simple roundhouses classification (Armit, 1991) - appear to
have been subsequently abandoned fairly early in their lives (eg
Bu of Navershaw, Pierowall, and possibly Calf of Eday). On one
level it has some logic - the effort involved in removing simple
but massive roundhouses to build a more refined tall broch may
well be a significant consideration. However it is a potentially
misleading approach to take this as the normal scenario: Howe
itself seems a clear exception, and to some extent Tofts Ness,
where big demolitions were accepted practice ahead of construction.
More significant though is the insight that Howe gives for the
range of existing and exposed broch sites, which despite their
appearance of being single-build structures, may conceal evidence
of multi-construction phases over a protracted duration.
This leads to a bigger question of why, if many brochs were brownfield
sites, it was desirable to maintain the site's use rather than
start from scratch. The on-site availability of reusable building
material has to be one practical advantage, though it is difficult
to see this as an absolute factor, particularly when building
a bigger or taller structure than its predecessor.
Instead the answers have to be rather more conceptual at this
stage, developing from the idea that the original site location
factors remained valid. These would include having a viable catchment
area, including perhaps fertile soils or sea/loch access in the
immediate vicinity, perhaps also elements of natural vantage,
and the specific requirement of a fresh water supply close at
hand, whether on the surface or underground. While these can be
analysed for any given broch site (eg Fojut 1982), the permutations
will be idiosyncratic and there will be a finite number of locations
in the landscape which combine them. These factors, though are
largely timeless in their desirability, and apply at a conceptual
level to earlier prehistoric eras. The What-If? Pixie can then
plant the idea that the landscape would have been developed through
preceding millennia, with many of the "good" locations
already captured, and the hinterland of these sites developed
with fertile soils being successively improved by human cultivation.
Thus it makes sense at a general level to reuse settlement locations
within a landscape rather than develop greenfield sites.
So land use patterns could conceptually remain fairly constant
through prehistory. This is too big a question to explore in this
paper, so will be left hanging beyond pointing out that the Bronze
Age paradox identified earlier in this paper does imply a continued
land use pattern bridging the Middle Bronze Age into the Iron
Age over a climatic hiatus.
All this is a long way from saying that all brochs were brownfield
sites. It is an inescapable impression from visiting a range of
broch sites in Orkney and elsewhere that many of them are in extreme
micro-locations, with highly exposed positions and/or very restricted
situations. The choice of these sites is usually regarded as being
dominated by defensive capabilities for a desired impregnability,
and within the terms of this discussion seem more likely to be
greenfield developments away from the established land use pattern
into more marginal situations. Ironically, a number of the brochs
already cited as having pre-broch site use also have these extreme
micro-locations (eg Borwick, Midhowe, Clickhimin), which illustrates
the dangers of applying glib generalisations to a set of complex
individual circumstances. The permutations are further widened
for broch groups in extremely close proximity (eg the trio centred
on Midhowe, the Keiss trio in Caithness), with a variety of extant
suggestions, usually based on socio-political grounds.
The greenfield or brownfield question is not capable of providing
a direct answer from our current knowledge level, but it should
be apparent from this discussion that there are dangers in assuming
greenfield construction for any given site without detailed investigation.
The brownfield scenario deserves further consideration, and opens
up many lines of enquiry which could give us a more viable context
to explain the broch phenomenon. In the interim, both possibilities
are legitimate considerations at the individual site level.
The duration of the broch period.
This is another aspect where further consideration may well be
overdue. The earliest dated broch-like structures so far found
date from c. 700 BC, all as it happens in Orkney (Bu of Navershaw,
Pierowall, Tofts Ness, possibly Howe Phase 3). The latest reliably
dated broch constructions occur some eight centuries later among
the southern outliers such as Leckie or Buchlyvie, with absolute
dating and secure Roman artefact typologies. Some brochs further
North, excavated before absolute dating became possible also show
Roman artefactual discoveries, all of which has led to a general
impression that brochs were being built over an 800 year span
to about 100 AD.
The difficulty is that we tend to assume that the picture we
now have of broch distributions, with anything up to 700 examples
known or suggested, represents a contemporary snapshot for some
undefined point during this span ie they were all in use together.
This assumption is worth unpicking, particularly here for the
early part of this 800-year span. If we assume that brochs were
being built throughout the period, it means that some sites which
eventually featured brochs were not broch sites at, for the sake
of argument, 300 BC which was the approximate mid-point of the
broch era. Therefore they were either sites without settlement
structures or they had some contemporary non-broch settlement
forms. The former permutation has a specific implication about
Iron Age population levels undergoing a massive increase during
the broch period from a very low starting point. This is difficult
to relate to the number of burnt mound sites in areas such as
Orkney and Shetland (even if - as argued earlier - the burnt mounds
are not the most convincing known settlement indicators for the
preceding period). The second permutation, that a significant
proportion of broch sites only developed brochs later on and had
preceding Iron Age site use avoids the extreme fluctuations in
population, but raises further issues about the rate of broch
proliferation and what enabled a given broch site to develop a
broch at the particular moment it did.
As broch construction is a substantial resource investment of
labour and materials (both in volume - stone - and scarcity -
wood), it is likely that some form of threshold capability had
to be crossed to enable construction. The threshold may well have
had economic, technological, social and political components in
a variety of permutations - differing over geographies and with
time - so again the approach is conceptual rather than absolute.
The insight does create a credible context in which a large proportion
of brochs would have been brownfield sites and where those sites
were most likely continually settled until the broch insertion
- the people become the constant with their needs and motivations,
not the building forms.
Implications for excavation strategies.
Throughout this paper, a running theme has been that we know
too little, despite the large number of broch sites that have
been exposed, explored and/or displayed. The evidence we have,
when assembled and interpreted is a long way from a convincing
explanation of the permutations of events represented by the brochs.
Howe produced a number of insights into what could be a more widespread
and more complex picture than previously evident, and one which
in my view has a better logic by opening up the explanations and
their frameworks. The Howe insights were derived however from
subtle pieces of evidence, scattered and very localised clues
which lacked an intrinsic coherence until the full picture was
assembled by the excavators.
At the 1998 Shetland broch conference, one plenary debate erupted
very forcibly with the plea that broch excavators have to target
the primary interior floor surface in their excavation if a useful
account of the site is to be constructed. This paper would extend
that further and suggest that excavators should also look at whatever
is immediately beneath the primary floor surface, and that they
should take every opportunity to explore underneath the broch
wall and to scrutinise the immediate vicinity of the broch structure.
The evidence will be very fragmentary, featuring isolated strands
without an immediate coherence, and will probably only come together
if a large expanse is taken down below the obvious structures.
This is a tall order, particularly if it involves removing key
structural features in an excavation for eventual display, but
is the only probable way of reaching these clues.
Whether it is realistic to hope for future broch excavations
on the scale of Howe is another question, particularly if the
Old Scatness strategy of stopping excavation at a higher level
for display is repeated to justify the excavation investment.
An alternative route might be to build these objectives into projects
on the number of sites which have already undergone exploration
and been left untended since (such as current work at Everley
and other sites in Caithness) or those which have been damaged
by marine erosion (which we are probably going to lose anyway
in the foreseeable future so the issues of destroying "good"
structures to investigate something that may or may not be present
become secondary). The Shorewatch scheme of coastal investigation
could be a very useful starting point here.
Part 4: conclusion
This article set out to explore some of the unusual findings
of the Howe excavation, particularly the apparent deviations from
the normal script, and to develop a perspective showing how these
findings could represent the tip of a previously unidentified
iceberg rather than a site which was simply an inconvenient aberration.
Hopefully, the clues highlighted from other sites give a new slant
on the issue of broch sites as potentially showing continued and
perhaps even continuous occupation of their chosen locations from
earlier periods, while also emphasising that the clues themselves
will be subtle and not always readily apparent.
In 1981, Noel Fojut posed the rhetorical question "Is Mousa
a broch?" as a device to highlight a number of architectural
differences at that site from the established model. His conclusion
was that Mousa definitely was a broch, but no other broch was
a Mousa. In glib parody of his approach, I suggest that while
Howe was definitely a broch (or series of brochs), a significant
number of other brochs could also turn out to be like Howe, if
only we start to look hard enough.
There is another issue which I am beginning to appreciate after
25 years of fascination with these structures. I started off,
as many others have, dazzled by their complexity and, in the few
surviving towers, by their magnificence. I wanted somehow to unravel
how such unparalleled architectures represented a unique phase
of Scottish prehistory, with a spectacular story to match the
evidence. Yet now, I feel the roots of the broch story are much
more mundane and easier to relate within the prehistoric context.
Brochs evolved from the context in which they originated - they
were not parachuted in as a fully formed package by an alien influence.
Some brochs are indeed massive, monumental and isolated, but many
others seem to fit unobtrusively into the landscape in a wide
variety of situations, whether comparing across or within parts
of the broch areas. The quality of stonework varies immensely,
as indeed does the stone itself and the architectural detail within
the concept of a massive and high round form, and the likely hinterlands
and therefore economic frameworks of the broch communities seem
to vary on a similar scope. Many of these comments echo recent
observations by Andrew Baines (2002, pp 13-16), and suggest that
we should no longer think of a broch phenomenon as a distinct
mystery, but that we should be trying to see the brochs as components
within a bigger context where all these differences can be fully
explored. We are moving away from a mindset - that brochs are
culturally diagnostic - which has underpinned much previous broch
study into a much more open perspective. This does not mean that
the broch story is diminished - the fascination is still there,
but its magic is in explaining how mundane roots maintained a
prolonged progression into such a magnificent final crescendo
and identifying the many permutations of context, and inspiration
which are represented. In fact, it is probably misleading to suggest
that there could be a single broch story - the reality is that
there are many intertwined paths to unravel within a framework
of social, geographical, technological and chronological variables.