Dear Mr. Hocknell,

For what it's worth, the Muslim-Croat Federation-Republika Srpska 
boundary has been reported by Mark Corson and Julian Minghi as 1,053 km 
('Political Geography of the Dayton Accords,' _Geopolitics and 
International Boundaries_, 1:1 (Summer 1996), 75-92; v. 83).  I'm 
assuming that they got this measurement from the US Government, and that, 
possibly, the figure represents a length based on the 1:50,000-scale 
sheets on which the boundary was initially depicted.

Which raises the point made by Galo Carrera as to the role of scale in 
the measurement of the length of a boundary or a coastline.  A related 
concept that might be raised is the purpose for which the measurement is 
to be used.  The pure definition of the length of a boundary would be the 
following of every turning and, in the case if rivers, meandering of the 
the line, with presumably as dense a netwok of turning points as possible 
(even though a true length would still be elusive).  This would certainly 
be the kind of figure relevant to such activities as boundary marker 
maintenance, border patrol assignments, etc.  But, what point is there in 
following every meander if the intent is to show the amount of 
"continental breadth" of a boundary?  For example, is the average person 
interested in the exact length of the Mexico-United States boundary, 
including every meander of the Rio Grande (where it has not been 
straightened out artificially), or in the distance the boundary 
"consumes" in traversing from the Gulf of Mexico to the Pacific Ocean?

This, in turn, can sound like just another manifestation of the scale, 
i.e., generalization problem, but it also raises the point of the 
practical comparison of a meandering river boundary and a straight-line 
or gently curving boundary.  Certainly, we would not want to generalize a 
boundary following the course of a stream that is flowing in very gentle 
flexures.  But should measurements of river boundaries be "straightened 
out," following perhaps the Thalweg of the valley (instead of that of the 
channel) when the stream channel begins to display pronounced meandering, 
and, at what point in the frequency of channel flexures should this 
begin?  How contorted should a riverine boundary be before we would 
consider abandoning measurement along the exact channel?

Another problem with following the exact channel of tightly meandering 
streams is that such streams usually exhibit frequent changes in course, 
rendering any measurement of length along them quite ephemeral.  This 
would be a problem, particularly in your case, if you wanted the exact 
length of the Palestine-Transjordan boundary along the Jordan River--not 
only because of the changing length of the Jordan's channel during the 
British Mandate period, but also in having to decide whether you want the 
boundary channel length during the 1920s or during the 1940s just prior 
to the end of the Mandate.  All of this, in turn, would depend upon 
the availability of survey sheets from either period upon which to base 
the measurements, and, especially, on the accuracy of those 
sheets (consider mapping methods in the 1920s).

If this notion of generalizing a winding river boundary seems outlandish 
or lacking in precedent, I would invite comparison with the determination 
of a "coastal front" in cases where proportionality of coastline 
frontages is instrumental to determining fair delimitation of maritime 
boundaries.  It clearly would be unfair to follow every promontory and 
inlet in measuring the coastline in a case where one state had a deeply 
indented coast and the other did not.

Sorry to wax so philosophical on you, but perhaps you might consider 
elaborating on the purpose of your measurements and whether you would 
want the exact length of say, the Palestine-Transjordan boundary along 
the Jordan River or an approximate Thalweg measurement.

Regards,

Brad Thomas


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%