3 Scope of the dataset
Neither archaeology nor archives have revealed a clue yet that waterships were ever constructed outside the region that is nowadays called the province of North-Holland. Hence the scope is limited to this region (fig. 7). The period in which the watership occurs spans at least four centuries. Its genesis and original appearance remains obscure since fourteenth and fifteenth century waterships did not show up yet in the archaeological record. Its demise, well known from archival information, is 1827 when the Dutch Navy decommissions the last watership from the service as a tug (Koningsberger & Oosting 1994; Boven & Hoving 2009, 55).
Figure 7 The present-day province of North-Holland in 1350. Each opening to the sea has a dam. The black lines represent the layout of the dyke system. (
, 41). The lakes are turned into land between 1612 and 1624.
The concept of a ship transporting live fish in a fish well is not unique to the watership, as it was already practised in Roman times (Boetto 2003). It is postulated that the first waterships were local vessels used in the many lakes and creeks of North-Holland, in the 14th century, to collect fish, transporting the catch to local markets and production centres. However it is not known what these ships looked like, since one of the oldest wrecks revealing relevant data is the ZN44 constructed somewhere around 1500. In Dutch archives the first scant reference to a watership is made in 1339. An English archive from 1420 indicates that it may have exported live fish across the North sea. Dutch archives refer to waterships in 1494 transporting live eel to Belgium and other provinces in the Dutch Republic (Ypma 1962, 44). Edam is mentioned as a town that owns a watership in 1462, equipped with trawl nets (Boschma-Aarnoudse 2003, 146 & 159). The archaeological and archival information is sufficiently convincing to identify the function of the watership in collection and transport. The export function of live fish to distant markets is not yet archaeologically proven, but likely (Ypma 1962, 33-36, 42, 44-45). This function, with a need to sail at sea and in coastal waters, may explain why the ship is built with a keel plank, S-shaped bottom fore and aft, and a strong lap-strake hull (fig 2). On the other hand many flat bottomed ship types, some with a fish well, are known to have sailed coastal waters and the North sea. This paper will not address the issues related to the genesis and original appearance of the watership, as adequate data are lacking.
The scope in this paper is the watership originating from North-Holland in the sixteenth and seventeenth century, since the dataset of 41 shipwrecks does not extend beyond this timeframe. This dataset was explored in search for a consistent and uniform basis that would allow for variables to be compared. It appeared that substantial reduction of data was needed, as well as a renewed effort to derive geometric measurements from field drawings in a uniform manner. Different reports show different gaps in the listing of geometric data, use different reference points, or mention no reference at all. For example ship length was measured or estimated in many different ways, depending on what was remaining of the wreck and the goal of the field investigation. Local differences in deposition processes account for differences in the way wrecks deteriorate. Sometimes a rudder or sternpost is all that remains, sometimes almost a full hull is available. Therefore consistency and uniformity in geometric data is always a stretch. Twenty-one wrecks were not sufficiently documented to meet these demands. They are also not available for additional retrieval of data in the future. However this limited dataset is not useless, as it still contains valuable qualitative information on construction details that is used in the assessment of change (table 1).
From seven wrecks only basic information is gathered, but there is potential for retrieval of information through additional fieldwork in the future (table 2). These wreck have only been explored in situ and covered again for preservation in situ.
Finally thirteen wrecks are sufficiently documented for the retrieval of comparative data in both a geometric and qualitative sense (table 3). These wrecks could be sequenced in time, based on an accuracy estimate of 25 years for the construction date of a new ship. The construction date estimates in table 3 are derived from several dating methods and from the appreciation that the life span of a ship must have been around 20 years. A seagoing ship had an operational life of fifteen years on average (Boschma-Aarnoudse 2003, 234). Two waterships served the Navy for 25 years between 1802-1826 (Acts 3161 and 3247, Notary Archive Monnickendam, nr 3584). The age of wreck ZN42-1 was over 20 years old when wrecked (Pedersen 1996, 67). The expected life of 20 years seems reasonable for a ship in the sixteenth century engaged in heavy duty trawling. This also means that the waterships in table 3 cover a period of eight to nine generations of newly constructed waterships. This should be enough to allow for change in ship building practices.
For the wrecks ZN44, OG33/34 and OU86 the original field data were consulted, as stored in the maritime archaeological archives of the Cultural Heritage Agency in Lelystad. For the other wrecks, listed in table 3, also well documented reports were used (van Gent 2002, 92; van Holk 1983; van Holk 1986, 20-21; van Holk & Immink in prep. a, b, c; Hulst & Vlek 1985; Pedersen 1996; Reinders 1986). The sintel typology method of Vlierman was used (Vlierman 1996).