You are on page 1of 30

Central Places and Major Roads in the Peloponnese

Author(s): G. D. R. Sanders and I. K. Whitbread


Source: The Annual of the British School at Athens, Vol. 85 (1990), pp. 333-361
Published by: British School at Athens
Stable URL: http://www.jstor.org/stable/30102854
Accessed: 05-12-2017 20:28 UTC

JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@jstor.org.

Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
http://about.jstor.org/terms

British School at Athens is collaborating with JSTOR to digitize, preserve and extend access
to The Annual of the British School at Athens

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE
PELOPONNESE

Route networks of past and present landscapes reflect patterns of h


social activity. By describing a network empirically and investigating
constituent points are, an idea of the relative importance of a given r
developed and measured. In the case of the Peloponnese, this appr
predicting which cities were of importance in the past. For most
history, the important centres are well documented and althoug
mathematical analyses tells us what is common knowledge, such
illuminate why certain places flourished when others did not. The results
central places but also how connectivity and accessibility, potentiall
variety of scales, drop off with geographical distance from a given centr
One of the most neglected monuments of Greece to have survived f
ancient and pre-modern road system. Although much useful wor
topographers such as Pritchett, Vanderpool and Hammond, to trace r
few, if any, analyses of the communications system as a whole.
fragments of roads still apparent on the surface a considerable bod
amassed. The travellers of the nineteenth century followed, as clos
ancient routes between the major cities of antiquity and were able to
accuracy, where Pausanias and other writers placed monuments, batt
They worked on the justifiable premise that contemporary line
through mountain passes and across fords were the same as in the
authors. Networks can be restored by means other than deduction
sionally, rare maps survive showing the major routes of the pas
Peutinger Table which records the Roman post roads for Greece giv
more important cities of the period. The Antonine Itinerary does no
but is useful for large sections of the Empire. A small number of m
routes of the early nineteenth century, a period before industrialisat
of Greece led to the creation of new settlements and lines of communication to serve new
economic impulses and the redevelopment of old settlements for romantic or nationalistic
reasons. Among these latter maps are George Gennadius' 1822 map of the Balkans, the
map in Leake's Travels in the Morea, that in the French Expedition de la Morie and General
Pelet's 'Carte de la Morde' made between 1829 and 1832.1
This paper sets out a method which allows the analyst to describe communications
networks empirically and is successful both in identifying routes of particular importance
and, from the linkages thereby portrayed, in identifying central places - thus proving the
adage 'all roads lead to Rome'. Specifically, it sets out to employ a method of analysis
borrowed directly from Social Geography, which in turn borrowed it from the mathe-
matical field of Geometry. The discussion is primarily concerned with the commu-

'G. Gennadius, Map of the Balkan Peninsula 1822; Expedition Scientifique de la Morde Paris 1834; General Pelet,
W.M. Leake, Travels in the Morea I 1830, Frontice piece; Carte de la Morie, Paris 1832.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
334 G.D.R. SANDERS and I.K. WHITBREAD

nications of the Peloponnese and relies, as far as possi


The success of the application suggests that it ma
networks, such as city plans and large non-domestic
divided into two parts; the first deals with the mathemat
second covers selected examples of actual networks th
the end, two BASIC programmes are furnished for tho
the analysis. The programme and instructions are avail
from the authors c/o The British School at Athens, 52

Graph theory and Network Analysis

Developed from a field of mathematics called topolog


and non-planar networks, has been used as a measure o
the early Ig6os, and the bibliography on the subject is
Graph Theory in the discipline of Economic Geograph
networks in an empirical manner. Garrison, in his st
Interstate Highway System in the Southeast United S
nodes to produce an index of connectivity (accessibilit
compared.3 Reed refined the methodology to demons
airline traffic networks.4 Although this approach is u
of networks and urban hierarchies, the data are relativ
nodes are weighted to reflect volumetric consideratio
measure, the S-I Index, has been suggested for the
accessibility of different networks.5
Those acquainted with transportation maps recognis
most familiar, a planar network, is one in which c
intersect at nodes or vertices (e.g. settlements and
non-planar network is one in which crossing vertices
example, airline route maps. Various indices of the
proposed. The simplest is the ratio (Beta Index) betwe
number of vertices (v) within the network (e/v). T
measure of connectivity and its values range from I
infinite number in non-planar networks.6 An equally
Kdnig Number, or the number of edges, in the shortest p
node furthest away measured in the number of links.
point 16 is 3 and for point 8 is 6. Low K6nig Numbers
within the network.7 The shape of a network can be
(delta), or the minimum number of edges between th

5 G.A. Jones,
2 P. Haggett, A.D. Cliff and A. Frey, Locational etin
Analysis al. 'Some Discrete
Graphs
Human Geography, Bristol, 1977, provide withintroduc-
a useful Applications to Region
tion to Graph Theory and extensive works'
basic bibliography.
Geographiska Annaler 52 (1970) 1
3 W.L. Garrison, 'Connectivity of the Interstate
6 Haggett, et al.High-
op. cit. 315.
way System', Papers and proceedings of7 K.J. Kansky, Science
the Regional 'Structure of Transpor
Association, 6 (1960) 121-137. tionships between Network Geometry
4 W.E. Reed, 'Indirect Connectivity and Hierarchies
acteristics', University ofof Chicago, Depart
Urban Dominance' Annals of the Association
Research Papers of American
84 (1963) 28-29.
Geographers, 60 (1970) 770-85. 8 Ibid. 21-23.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 335

A9 ,17,
8N I
210 N 18
'6 81 -I9 ,10
1= "8 I ,1 2 K ,12 A - .20.
I 7.. ,s: .11 111 j 19N
I
- N
12 13 21
2' ,4
I
1/ .7- .16N N K.
3- :7: I4
I2
I 2 I I 14. 22
2i"
C3/
30 I
'3<9 ,15 2 28
iI
30 27 25
~29/
-26'
1 2 3

FIG. i. Hypothetical networks and their connectivity.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
336 G.D.R. SANDERS and I.K. WHITBREAD

most sensitive measures of connectivity, and the one considere


Shortest Path Distribution. This is a matrix of the shortest path,
between every pair of vertices in the network; the sum of the
single given vertex to all others is a reasonable measure of the c
relative to the others within the network.
An example of the value of Binary Shortest Path Distribution (BSP) and how to
calculate it is easily provided. Figure I.I is a hypothetical road network connecting seven
settlements by twelve routes. This can be represented in a binary connectivity matrix by

application
if of the
there is none. formula
The binary Ai = I ifof
matrix there is a direct link
the hypothetical between
system vertices in
represented i and j, or
figure I.IAis
represented in Table I.

M' I
I

2 1-1000o13

3 oI I oo1 3
4 1013
5 ooo
6 I0o
7 6

Beta Index = 12
diameter = 2

Table I

The sum (I) of each row gives the number of single step links possible from each no
and, in this example, demonstrates that vertex 7 is the most connective and accessib
while vertices I, 2, 3, 4, 5, 6 are least connective. In more complex networks t
connectivity is not so simply established and thus the relationship between nodes no
directly linked must be derived. This can be achieved by powering the matrix (multiply
the matrix by itself) until there are no zero off-axis values remaining. From this th
Shortest Path Matrix can be created.
The mathematics of matrix multiplication are simple but long and repetitive, espe
for large matrices. The original binary matrix, M' (Table i), is multiplied by its
produce its square, M2 (Table 2). M2 can then by multiplied by M' to calculate M3 a
on. In order to multiply two matrices, the elements in the rows of M' should be mult
by the corresponding elements in the columns of the other matrix and then
together. For example, the value of M2 [row I, column I] is (oXo) + (IxI) + (oX
(oxo) + (oxo) + (I x Ix) + (I x I) = 3 and for M2 [2, 3] is (IXo) + (oX I) + (IXo) + (
+ (oXo) + (oXo) + (IX I) = I. M2, (see Table 2) contains the total number of two
links in the network including those originating and terminating at the same point.
The initial Shortest Path Matrix (SPM') is identical with the original binary matrix
as it documents the existence of single step links between nodes. SPM2 will sho
presence of both one and two step links. It can be created by comparing SPM' with
Most of the two step links in M2 are false in that they retrace their steps (for example; M

[I,I] =3 i.e.
because two step
a one stepspath
frombetween
I--2-I, I--6---I
nodes and I-7---I)
already or they
exists (M2 are irrelevant
[1,2] =I, a twofor SPM
step

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 337

I-7--)2 where the one step link I-2 is already recorded in M'). If a value appea
off-axis element in M2 where no value exists in SPM', the number 2 (i.e. the
which M has been raised) can be inserted in the appropriate slot in SPM2. Thi
that a two step link exists between the two relevant nodes. This process cont
there are no off-axis zero values; it will be noted that the powering continues t
of the diameter of the network. Fortunately, the mathematics are greatly sim
using an algorithm, in this case, for a given node Nj
dik:=min (dik, dij + djk) where i()k()j
[where := means replaced by, min means minimum, () means not equal to, i m
number, k means column number and j means source node.]9
M' i 234567
I - I O O O I I

2 I - I O O O I

3 o i - Io o I
4 o o I- Io0
5 o00oo -II

7 I O I - I-

M 2 I 2 34 5 6 7 M2(MlxM1) SPM2 X Rank


I - IOOOII 3 121213 I -12221192=
2 I - I 0 O 0 I I 3 2 I 2 3 2 I - i 2 2 2 I 9 2=
3 O I I 0 I 2 I 3 I 2 I 3 3 2 1 - I 2 2 I 9 2=
4 oo001- 101 I 213123 4 221- 12192=
5 o0o 0 - I I 2 I 2 1 3 I 3 5 2221- I 1 9 2=
6 I o o o I - I I 2 I 2 I 3 3 6 1 2 2 2 1 - I 9 2=
7 I I I I I I - 3 3 3 3 3 3 6 7 I I I I I I - 6 I

Table 2

Weighted Shortest Path Matrices (WSP) can also be calculated. Instead of using bin
notations to indicate merely the presence or absence of a direct link, a value, for exam
cost, quantity, kilometres or hours, between the linked nodes is inserted. Thus, if the valu
between points 2 and 6 is four hours travelled or four pots exchanged, then 4 is ente
instead of the I using the binary matrix. Points not directly connected are, in
calculation, considered to have an infinite value. The advantage of using weig
matrices is that the highest and lowest value paths across the network can be calcu
and a more refined measure of connectivity and interaction can be derived. Allowance
also be made for unequal movement in opposite directions since, for example, the vol
moving from i to j may not be the same as that moving from j to i. This can be expressed
the algorithm
dik:= max{dik, min(dij, djk)}.10
By systematically removing links or vertices within a network and recalculating
solution matrix, the impact on the whole network caused by the removal can

9 T.C. Hu, Integer Planning and Network Flows. Reading,


1972 P.49ff.
Mass. 1969 p. 157. For further reading see R.B. Potts and10 T.C. Hu, op. cit. I55.
R.M. Oliver, Flows in Transportation Networks, New York

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
338 G.D.R. SANDERS and I.K. WHITBREAD

established and is a good measure of that point or route's importanc


value of this exercise should be appreciated if, for example, one consid
transport patterns in San Francisco occasioned by the collapse of the
in the 1989 earthquake.
The degree to which connectivity within a network is altered b
measured and is illustrated in the second example (Figure 1.2, Ta
hexagons sharing two common vertices form a network of twelve vertice
edges. The Beta Index is 1.91 reflecting a greater degree of connectiv
than the first example, and the diameter is 4. The effect of add
hexagonal network illustrated in the first example is to displace the
connective point. Applied to an even more complex network (Fig. 1.3,
hierarchy of relative connectivity can be calculated. The seven adjace
edges and 31 vertices can be described as a 31 x 31 matrix: the Beta In
diameter 6. The Shortest Path Matrix sums show five distinct levels
centre point, number 16, being the most connective and the points at the
least connective."

Fig. 1.1 Figure 1.2 Figure 1.3 Figure 1.1 Figure 1.2 Figure 1.3
No. 1 Rank X Rank X Rank No. 1 Rank I Rank 1

1 9 2= 22 3= 104 5= 16- - - - 66 1
2 9 2= 26 4= 104 5= 17- - - - 104 5=
3 9 2= 26 4= 88 4= 18- - - - 104 5=
4 9 2= 22 3= 71 2= 19- - - - 88 4=
5 9 2= 17 1= 71 2= 20- - - - 82 3=
6 9 2= 17 1= 88 4= 21- - - - 104 5=
7 6 1 18 2= 82 3= 22- - - - 104 5=
8 - - 22 3= 104 5= 23- - - - 88 4=
9 - - 26 4= 104 5= 24- - - - 82 3=
10 - - 26 4= 88 4= 25- - - - 104 5=
11 - - 22 3= 71 2= 26- - - - 104 5=
12 - - 18 2= 82 3= 27- - - - 88 4=
13 - - - - 71 2= 28- - - - 82 3=
14 - - - - 71 2= 29- - - - 104 5=
15 - - - - 71 2= 30- - - - 104 5=
31- - - - 82 3=

Table 3. X is the row sum (Shortest Pa

SOME ARCHAEOLOGICAL APPLICATIONS OF THIS METHOD TO GREECE

A. THE GREEK SECTION OF THE PEUTINGER TABLE

A suitable test of Graph Theory applied to ancient networks would, ideally, be one similar
to those used by economic geographers; a network with nodes and links already

" For theories of distance-decay rates and centrality see Funktionen, Jena (I933). Note how accessibility declines
J.H. von Thiinen, Der Isolierte Staat in Beziehung auf Land- with distance from the central point and how points 7, 12,
wirtschaft und Nationaldkonomie Hamburg (1826) and W. 20, 24, 28 and 31, each two links from node 16, exert their
Christaller Die zentralen Orte in Siiddeutschland: Eine own pull whereas 3, 6, io, i9, 23 and 27, also two links
iikonomisch-geographische Untersuchung iiber die Gesetzmiissigkeit from node 16, have a lower rank.
der Verbreitung und Entwicklung der Siedlungen mit stiidtischen

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
v

20
121
15
19

8)

18i 17'
11
27

>9
7

16

1< 2

S2,
10
13

i29
25 22
23
0

FIG. 2. The Peutinger Table for Greece: The Peloponnese and Achaea. For identification of Peloponnesian cities, see
Table 4.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
340 G.D.R. SANDERS and I.K. WHITBREAD

documented rather than chosen by the analyst, one with s


nodes to check the results, and one that has few externa
these criteria is the ancient map of the Peloponnese depict
The Peutinger Table (Codex Vindobonensis, 324) is a
parchment, 6.82 metres long by o.34 metres wide, proba
Augustan prototype and subsequently modified on a num
thought to show roads and stations used by the imperial
network of primary roads used for government busines
centres of the Roman world via a series of towns and min
network is comparable to the American Interstate sy
pioneering study of Geographical applications of Graph Th
The Greek section of the Peutinger Table (Figure 2) was
Miller'" and his identifications of sites has been updated
the purposes of the test, Pritchett's identifications of rou
in almost every respect. However, the route from Me
Pritchett is the direct one over Taygetos through the La
to timings estimated by I9th century travellers, was slow
as 19.5 hours at walking pace.'6 The alternate route, throu
the north end of Taygetos and down the Eurotas valley, is
a total of 15.15 hours at walking pace.'7 It is probable th
Messene to Sparta fell in with the Megalopolis to Sparta
Leondari.'8 Whether or not the ancient road from Sp
followed the modern road is questionable. There is goo
following the Eurotas from Sparta, crossing the river at
towards the Helos plain.'9 There is a distinct possibili
Peristerion (formerly Tsasi) and that separate roads le
Asopos. For the purpose of the analysis it is assumed t
crossing the hills of Bardounia on a direct line to Gytheio
road continued to Asopos.
The modified map of the Peloponnese from the Peutinge
2. It can be seen that the network is almost entirely enc
except via the Isthmus and by sea. The archaeology of
documented and few, if any, of the sites appearing on th
identified either positively or by speculation. A brief disc
Pritchett.20 The area is therefore well documented for the pe
The passes of the central Peloponnese have always bee
across the area. By closing these passes, a strong state was
around the periphery and thus severely impede the rate

12 W.H. Stahl, 'By their Maps youandshall Know


W.M. Leake, them'
Travels in the Morea London (1830) 366-7
Archaeology 8 (1955) 152-3. for Thuria to Mavromati.
13 A.H.M. Jones, The Later Roman Empire II op.
17 Gell, Oxford
cit. 6o-i, (1964)
74- 6, 213-2o.
83o-834. 18 W.K. Pritchett, Studies in Ancient Greek Topography V Los
14 K. Miller, Itineraria Romana, Stuttgart (I916) 559-571. Angeles 1985, 61-68 makes the case for the existence of
15 W.K. Pritchett, Studies in Ancient Topography, III Roads this route from Sparta to the Stenyklaros plain.
pp. 97-288. 19 Leake, op. cit. I 194-5.
6 W. Gell, Itinerary of the Morea London 1817, 234 for 20 Pritchett, III 237-288.
Mistra to Kalamata; idem. 65-6 for Thuria to Kalamata

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 341

commodities travelled from point to point. In the fourth century Megalopoli


important centre. After Leuktra Epaminondas lent Theban troops to secure Ar
the city was founded as a synoecism of forty small Arcadian cities.21 Cir
however, Kleomenes demolished and burnt the city22 and, despite subseq
cupation, it was 'deserted' by the time of Strabo.23 Pausanias laments the
grandeur and the unfulfilled hopes of its founders, but the settlement co
function at least into the fourth century AC.24 On the other hand, Megalopolis
Tegea continued to flourish under the Romans25 and even on down to until t
Byzantine period.26
In contrast to Megalopolis, Corinth and Patras were the great cities in the reg
the Empire. Corinth was reduced to ruins by Mummius in 146 BC but was no
deserted.27 The city was refounded as the colony Laus Julia Corinthiensis by Caesar
becoming the capital of the province of Achaea. It was served by two harbou
Kenchreai on the Saronic Gulf and the other at Lechaion on the Corinthian Gulf. After
Actium in 31 BC, Patras was refounded by Augustus as his colony Aroe Augusta Patrensis and
was declared civitas libera. To judge from the wealth of monuments, Patras was one of the
primary cities of the Peloponnese.28 The city stood close to the narrowest part of the Gulf
of Corinth controlling the straits much as Corinth controlled the Isthmus. Its location
afforded an easy and short crossing to Naupaktos providing the city with good commu-
nications with Western Greece.
Places on the Peutinger Table are represented by a variety of symbols. Most nodes ar
shown by a kink in the road labelled with the name of that stage. Over 500 places ar
singled out for depiction by vignettes. The most elaborate, pictures of enthroned emper
or personifications of the cities themselves, are reserved for Rome, Antioch and Consta
nople, while Nicomedia, Ancyra, Aquileia, Nicaea, Thessalonica and Ravenna are ea
shown as walled cities with numerous towers. Selected baths, horrea, harbours and tem
also make an appearance but the vast majority of vignettes are simple two-towe
symbols of various forms. For Achaea, the Peutinger Table has ten of these latt
vignettes, at Athens, Megara, Corinth, Argos, Epidauros, Boiai, Methone, Olympia an
Patras. These vignettes seem to have had some significance and possibly identified ci
of extra importance.29
Four different tests were run for the Greek section of the Peutinger Table and ar
presented below. The first of these analysed the roads of the Peloponnese alone a
although the results were incompatible with the archaeological record for the Roman
period, proved of interest for the fourth century. The remaining three tests were designed
to discover why the first test apparently failed and in doing so produced valuable resul
The final test, an analysis of roads and notional sea links in the province of Ach
produced data which closely replicated the expected pattern.

21 Paus. VIII.27.4. 26-37. The politea of Corinth was ager publicus and conse-
22 Plut. Kleo. 25, Paus. VIII.27.16. quently was worked. The working population and over-
23 Strabo 8.8.I, 16.1.5. seers evidently lived in Corinth.
24 For a brief history see E.A. Gardner, et al. Excavations 28 F. Petsas 'Patrai' in R. Stillwell ed. The Princeton
at Megalopolis, London I892, 1-5. Encyclopedia of Classical Sites, New Jersey 1976 86I-2.
25 Paus. 8.45-54. 29 Stahl, op. cit. 153-5. A.C. Levi and B. Trell 'An
26 A. Bon La Moree Franque Paris (1969) 522-25. Ancient Tourist Map' Archaeology 17 (1964) 227-31 how-
27 C.K. Williams, 'The Refounding of Corinth: Some ever, consider that the vignettes show facilities available to
Religious Attitudes' in S. Macready and F.H. Thompson,travellers and were not used to identify important cities.
eds. Roman Architecture in the Greek World, London, 1987

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
342 G.D.R. SANDERS and I.K. WHITBREAD

Analysis of Peutinger Table road links for the Peloponnese

The graph of the Peloponnese section of the Peutinger Table is


It consists of thirty one nodes connected by a total of only th
I.o97). Argo-corinthia is well served by routes and has a high
contrast, the remainder of the Peloponnese is covered by two
Megalopolis and Olympia and by a branch running from
Laconian Gulf. Viewed subjectively, the network suggests, fr
settlements and communications, that Argo-corinthia is the m
the Peloponnese and that Megalopolis, Olympia and Leonda
importance. Moreover, Boiai appears to be geographically remo
The results (Table 4.') show that, contrary to expectation
connective node and that a corridor of high connectivity exten
to Megalopolis and from there to Melaineiai and Leond
connectivity can be discerned as in the hypothetical cases out
cluster of values around the mean value of I55.81. 55% of the v
I55+/- 20 while the remainder extend from iii for Megalo
pattern of connectivity produced (Fig. 3.2) shows that the stro
by Megalopolis is affected by the local pull of Argo-corinthia an
of Olympia and Leondari. Megalopolis may, in terms of the P
considered to be the central place, Argos and Corinth seen
Olympia and Leondari as secondary but less important places.
scale, Boiai, Asopos and Pylos are all relatively remote.
The computed pattern does not, in this example, reflect arch
larly well - Megalopolis is highly ranked while Corinth and P
respectively, are relatively remote. Nor do those apparent
coincide with those singled out in the Peutinger Table for depi
Corinth, Argos, Epidauros, Sparta, Boiai, Methone, Olympi
the pattern produced does illustrate the economic and politic
century when the city of Megalopolis was founded. Evidently
considerably and the accessibility by land routes, which had go
no longer vital under the Romans. One possible reason for th
expected and observed data may be that the distance between
Alternatively, the road network may play only a secondary r
network, and sea links may have been more important in
importance of cities.

Travel time for the Peutinger Table Peloponnese road links

Using data derived from travellers' diaries, the time travel ma


Table 4). Leake, Gell and Finlay each give detailed timings, m
for every stage of their respective tours. Since it was their origin
the known sites as possible, their records offer a reasonably a
long movement between major cities took and what routes w
three travelled on horseback but were guided by a man on fo
by a baggage train. The speed of the guide was the controllin
was sent on ahead or allowed to arrive late. Between the thre

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 343

1. PELOPONNESE 2. PELOPONNESE 3. ACHAEA 4. PELOPONNESE-


+ SEA LINKS + SEA LINKS TIME
(Hours)

PT Name Rank I ,/N-1 Rank 1/N-1 Rank I _/N-1 Rank , /N-1

1 Megalopili 1 111 3.70 1 110 3.67 20 227 4.93 1 653.5 21.8


2 Tegeas 2 115 3.83 2= 111 3.70 16 219 4.76 3 675.2 22.5
3 Argos * 3 119 3.97 2= 111 3.70 8= 198 4.30 4 699.2 23.3
4 Melena 4 123 4.10 8= 119 3.97 23 242 5.26 7 764.8 25.5
5 Leondari 5 124 4.13 15 127 4.23 28 258 5.61 2 673.0 22.4
6 Nemea 6 127 4.23 5= 117 3.90 11 205 4.46 6 740.0 24.7
7 Olympia * 7 133 4.43 10= 121 4.03 22 241 5.24 12 836.7 27.9
8 Sicione 8 137 5.57 8= 119 3.97 6 195 4.24 8 778.7 26.0
9 Micenis 9 138 4.60 13= 123 4.10 12= 211 4.59 5 720.0 24.0
10 Messene 10 145 4.83 26 148 4.93 29 271 5.89 13 840.3 28.0
11 Cleonae 11 146 4.87 10= 121 4.03 7 197 4.28 10 792.0 26.4
12 Netide (Elis) 12= 147 4.90 13= 123 4.10 18= 226 4.91 20 964.3 32.1
13 Lacede- 12= 147 4.90 27 149 4.97 31 282 6.13 14 846.2 28.2
mone

14 Epitauro * 14 148 4.93 21 137 4.57 8= 198 4.30 18 904.3 30.1


15 Agira 15 152 5.07 16 128 4.27 12= 211 4.58 17 896.2 29.9
16 Samacos 16 154 5.13 23 144 4.80 30 276 6.00 19 922.7 30.8
17 Corintho * 17= 156 5.20 7 118 3.93 1 183 3.98 9 791.3 26.4
18 Cyllene 17= 156 5.20 4 116 3.87 8= 198 4.30 22 1004.8 33.5
19 Dyme 19 162 5.40 24 146 4.87 14 215 4.67 28 1101.8 36.7
20 Agion 20 163 5.43 18 132 4.40 15 216 4.70 21 993.5 33.1
21 Pathras * 21= 166 5.53 5= 117 3.90 2= 186 4.04 24 1055.3 35.2
22 Asine 21= 166 5.53 28= 155 5.17 26 254 5.52 23 1023.6 34.1
23 Gytmon 23 172 5.73 30 161 5.37 25 250 5.43 29 1133.8 37.8
24 Cyparissa 24 175 5.83 22 142 4.73 27 254 5.52 25 1061.5 35.4
25 Mothone * 25 178 5.93 28= 155 5.17 21 240 5.22 26 1095.7 36.5
26 Istamo 26= 185 6.17 25 147 4.90 5 191 4.15 15= 854.7 28.5
27 Cencris 26= 185 6.17 19= 134 4.47 4 190 4.13 15= 854.7 28.5
28 Lechi 26= 185 6.17 12 122 4.07 2= 186 4.04 11 811.7 27.1
29 Pylios 29 188 6.27 17 130 4.33 17 220 4.78 27 1099.0 36.6
30 Asopos 30 199 6.63 31 166 5.53 24 245 5.33 30 1370.5 45.7
31 Boas * 31 228 7.60 19= 134 4.47 18= 226 4.91 31 1571.3 52.4

Table 4. The Greek section of the Peutinger Table. Row sums (


based on road links in the Peloponnese, road and sea links in t
and time travel in the Peloponnese respectively. * denotes nodes

was covered in depth. Leake and Gell travelled exte


but did not cover the eastern seaboard. During his m
the whole area exhaustively but lost the notes of hi
publish it. The surviving fragment is, fortuitously
precisely those parts not fully explored by Gell and
The routes taken by the travellers followed, f
communication dictated by the relief of the Pelopon

30 G.D.R. and J.M, Sanders, 'George Finlay in Laconia


and Arcadia' Philolakon forthcoming.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
344 G.D.R. SANDERS and I.K. WHITBREAD

High connectivity
(111-127)
'Peutinger Table'Vignette'
1. Low connectivity 2.
(185- 228)

Sea Sea
route' route

High (110-118)

connectivity
High connectivity
(183-200)
Mloderate connectivity 3. Moderate (126-137) I
(201-230)

FIG. 3. Connectivity diagrams of the Roman Peloponnese. i. Vignettes marked in the Peutinger Ta
Peloponnesian roads, 3. Analysis of roads and sea links of Achaea, 4. Analysis of roads and sea links of th

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 345

High connectivity
(EZ=3921-4440x10 mins.)

Moderate connectivity
(E= 4441-5534x10 mins.)

FIG. 4. Connectivity of the Peloponnese road network measured in time.

taine alluvial plains via passes, followed the course of rivers, crossing them at anci
bridging points and fords and, where possible, skirted the difficult interior by follo
the coastal plains. Occasionally the curious antiquarian left the direct track and trav
cross-country to visit a site not directly on his path. The routes themselves can ha
changed little since antiquity although concentrations of settlement may have mov
few kilometres for one of any number of reasons.31 Similarly the importance of va
routes has changed through time in response to the development of the settlem
linked.
The Time Travel Matrix contains the time taken to travel from each node to all the
other nodes in the network. The index of connectivity is, in this case, the sum of t
of time figures, that is the time taken to travel from any one point to all the othe
weighting factor time is primitive, but is at least a human dimension. Giv
unevenness of the survival of archaeological data, time is probably one of th
complete bodies of information available and is of practical use for other purpos
pattern which emerges duplicates the pattern above in many respects. The nodes
in time to all others are, in order of rank, Megalopolis, Leondari, Tegea, Argos, M
and Nemea. Certain towns are temporally 'closer' than their rank in the test

31 Pritchett, III, 54-Io0 traces the ancient road from


minor details, it follows the same route as that taken by
Argos to Hysiae and from Tegea to Hysiae. Except Leake
for and others.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
346 G.D.R. SANDERS and I.K. WHITBREAD

(compare Table 4.4 with 4.1I) such as Corinth, Lechaion and Ken
apparent hierarchy and, unlike the first test, the range of values is even
Asopos and Boiai which are exceptionally remote. Figure 4 show
relatively rapid communications stretching from the Isthmus to Ol
Sparta. The result is informative about the transport network and th
those expected, but it does not account for the discrepancies betwe
evidence and the first analysis. In the Roman period, then, neither ro
time seem to be major factors, yet locally remoteness from other l
important; in both analyses Patras is ranked low despite its pre-emin
useful product of the time analysis is the solution matrix itself (Tabl
the travel time by foot between different places; for example, it wou
42 hours to walk from Mycenae to Pylos, in other words, about five days

Analysis of Peutinger Table road and sea links for the Peloponnese

The role of sea transport is difficult to gauge without specific


transportation. Sparse coverage is given for the Aegean in the Itiner
which distances are recorded in stades for some but not all entries - from the Isthmus to
Naupaktos is 750 stades.32 The Peutinger Table for the Peloponnese does have one sea
link marked shown as Traiectus Stadiorum. This was supposed by Miller and Pritchett to be a
sea link between Boiai and Kythera.33 Theoretically, a harbour town has almost limitless
connectivity with other harbours, the sea presenting few constraints to the passage of
ships. Unlike land transport maps which are planar, sea transport maps are effectively
non-planar. For the sake of testing the degree to which sea passages affected the
connectivity of nodes, links were drawn between adjacent harbours of some standing, such
as Lechaion, Gytheion and Kyllene. A number of ports have been assumed, for instance
Argos with its excellent bay probably had facilities, but for places like Aigion, the
possibility has been suppressed. Long distance connections have been ignored deliber-
ately, largely because such links are difficult to control in a test situation. The link
between Boiai and Kythera was omitted to maintain the number of nodes at 31 and to
facilitate comparison with the initial test.
The new network has 45 links and a Beta Index of 1.45 reflecting a much more
developed system than the simple road system analysis above. The range of values is less
than half of the original BPM range, extending from I Io to 166 with a mean of 13I.7 +/-
15.6. A hierarchy of connectivity can be recognised: a group of three sites including
Megalopolis, Tegea and Argos stand alone; a second group with values from I16 to I23; a
third with a spread from 127 to 132; a fourth from 142 to 149 and, finally, an incohesive
group from 155 to 166. A broad analogy can be drawn between this hierarchy and that
shown in the third hypothetical hexagonal example above. From Table 4 it can be seen
that, as in the first test, Megalopolis, Tegea and Argos are still highly connective, but the
differences in values between them has sharply declined. Almost as highly valued as these
three are cities such as Corinth, Patras and Kyllene, while Pylos and Boiai are no longer
the remote places they seemed in the first example.
By weighting the network to include sea connections the results reasonably coincide

32 K. Miller, op. cit. LXVII-III. 257-8.


33 Ibid. 569, fig. 182; Pritchett, Ancient Topography III

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 347

Time=Mnuts/Io

TABLEOFIMRVD

123456789 0 2 3 1

1Isthmia0263785941208365497120 2Kenchrai6013785942 0183654927 01 3Corinth104527896 312586071 4Lechaion1709253864 021758 5Kleonai28190374526138079 6Nema32780951 42768301 59 7Mycena5034126975318209451230894 8Epidauros1263795016237 05932641807956 9Argos61485230 967214083 214089 10Tega968753421905687421 1Mcg/polis4738965012489735612 12Leondari56437809124536701285943 13Spart2089715402 167948021 67930218 14Asop3209576142093872415630 15Boia368940721653479286034 16Gytheion273045 216750921684073521 17Mesn2089354127608 93271 18Asine254930761249503871 2963 19Methon287605314 280793642 20Pylos3189725041 932576041 96720843 1 21Kyp/sia9860724319560287 1435096281347 2Samkos65 37419208 74105326894 23Olympia4180953214789503621 24Mel/nai198563 70215864972031 25Elis79601483 5260179384 5 26Kylen350412795603281490623147 27Dyme194852064 379152068913702456 28Patrs17056 42893057 642190358 29Aigon1078352609473251647089 30Aigera8976145230967418053 31Sikyon829406735182946350 TABLE5.ousethabl,findmr pceoign,tharlumdceownt hrfquied destinao.Thfgurvxpedinmts/Io.Eal:Myc(#7)P20is51xIO,.emnutawlkgpci41hors and50miutesor45daysitnblad.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
348 G.D.R. SANDERS and I.K. WHITBREAD

with the subjective and archaeological expectations. Argos, Patras an


defined as centres of some importance. A reasonable conclusion
combination of road and sea connections, in the case of the Pelopon
role in the determination of central places.

Analysis of Peutinger Table road and sea links for Achaea

So far, only the Peloponnese in the Roman Imperial period has been
time, the Peloponnese was not a self-contained unit when considere
than geographical terms; it was merely the southern portion of the
Achaea and, under the Tetrarchs, the Diocese of Achaea. In consider
of the smaller portion, the larger whole should also be examined.
The province of Achaea in the second century extended as far
Mountains (Fig. 2). From the Corinthian Gulf, the border followed
far as Karpenision and thence followed the Sperchenision river t
section of the Peutinger Table for this portion of Greece has 46 no
links including hypothesised sea links as well as the route betwee
(Beta Index 1.48). Central Greece was served by a large loop with ro
Nikopolis, Larisa and Athens but the network for the northern porti
of two long branches emanating from Eleusis and Megara. The loop
Peloponnesian network by a single road across the Isthmus and by hy
between Megara and Kenchreai, Kreusis and Lechaion and Naupaktos
the map, one would expect Corinth to be the central and most conne
sub-centres within the Peloponnesian network have already been id
WPM analysis above (test three) and should exist at Argos, Boiae, Ky
The computed connectivity (Table 4, Fig. 3.3) for Achaea shows a r
pattern for the Peloponnese than that hitherto seen; the most connec
whole area (including Central Greece) is Corinth from which corrid
tivity extend outwards towards Megara and Aigira. In terms of con
Lechaion are almost as highly ranked as Corinth while Argos, Epida
and Pylos all have high values relative to their neighbours. Not only
square with the expected archaeological and historical data but it als
with the hierarchy presented by the Peutinger Table itself. Of the
guished by a vignette on the Peutinger Table section for the Pelopon
Corinth, Patras, Argos, Boiai and Epidauros either have a very high c
rank highly in comparison to their neighbours. It is conceivable th
Methone was misplaced and actually belongs to Pylos which, like Bo
peak connectivity value. The relatively low values for the remain
Olympia must be explained in another manner. Olympia was a
religious and athletic centre which should account for its vignette.
was cut off from the west and east by a high mountain barriers and
Gytheion. Its very remoteness in the network and its history made i
for the administrative centre for its fertile hinterland.
Kyllene is something of a paradox. Its high connectivity value is not
the archaeological or historical record and yet, on reflection, is
offered a convenient landfall for mariners sailing the trade routes of
and to merchantmen approaching the Gulf of Corinth. It was the on

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 349

the extensive plain of Elis and must have handled a considerable volume of agricultur
produce for export. In the Frankish period Kyllene, at that time Clarenza, was the m
harbour of the Morea and was the fortified port of Andravida, the capital of
Principality. It supported the mint of the Principality and, later, of the Acciaiuoli, a
Florentine banking family. The archaeological remains are quite impressive. From
photographs the Roman city grid can be traced to the west of the Medieval fortificati
and the moles and basin of a large excavated harbour can be seen to the east. On
ground, there is a dense scatter of pottery dating from the Classical to the Late Rom
period. Evidently Kyllene's importance has escaped the notice of archaeologists a
historians.
There can be little doubt that this analytical method can predict hypothetical central
places. It can also be used with real data for the same purpose with remarkable success.
Failure to match expected with actual results in this example was a boon rather than a
hindrance since it provoked a number of important questions about what did determine
centrality in the Peloponnese. Had the null hypothesis been 'The road network of the
Peloponnese section of the Peutinger Table will predict the most connective/central
places of the late Classical and Hellenistic Period', the results of the initial analysis would
have been reasonably satisfactory. Instead, further tests showed that for the Roman
Imperial Period; i. the road system alone was not indicative of relative centrality; 2.
distance in terms of travel time was not a direct constraint on accessibility except, perhaps
in the case of Sparta; 3. sea communications together with road links either played a large
part in the importance of central places; 4. although a point may have little significance in
a small area, its importance within a larger network can be supreme for example, Corinth
and Patras, and 5. the vignettes on the Peutinger Table do seem to refer to important cities
or stations. Furthermore, a useful chart of travel timings by foot between cities was
generated.

B. THE PELOPONNESE IN 1822

A portion of George Gennadius' 1822 map of the Balkan Peninsula provides a usef
comparison to the Peutinger Table (Fig. 5). 34 The map shows the whole of the Balkan
peninsula with the principal towns and roads marked, and reflects the communication
network at a time when transportation methods were still primitive and roads sti
rudimentary.
The network for the Peloponnese is better developed than that illustrated by the
Peutinger Table. It has 33 nodes connected by 44 links (compare the Beta Index of 1.333
with that of I.og97 for the Peutinger Table). The basic network, however, is similar, with
routes following the same paths between cities located, more or less, in the same positions.
A noticeable difference, however, lies in that the focus of roads and nodes has diminished
in Argo-corinthia but has increased radically in Messenia. Viewed subjectively, the
expected result is that the central point of the Peloponnese will have shifted slightly from
the Argos to Megalopolis axis to concentrate on the Karitaina, Tripolis and Leondari
nodes.

In 1770 Tripolis became the capital of the Morea and had a population of circa Io,ooo

"4 Father of the bibliophile after whom the Athenian


library was named.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
350 G.D.R. SANDERS and I.K. WHITBREAD

25
23#
24
~22
t~9
-~t:
28'
-A
21 'T27

.20~C i

J3
19 >0
h
75
32 6
\33
9"
~e~
l8,! Ilk

~4/
8Y

15'M
1~42

16~Cj17
9<

FIG. 5. George Gennadius' 1822 map, Peloponnese section. For identification of nodes, see Table 6.

living in some 2,500 houses.35 It was the administrative centre of the whole area and the
residence of the Beglerbeg.36 As a result the postal roads from Patras, Gastouni, Arcadia,
Navarino, Modon, Corone, Mistra and Corinth all had their termini at Tripolis.37 Patras
was the second city of the Peloponnese in the early nineteenth century. It was of a
comparable size to Tripolis with as many as 2,500 families. As in the Roman period, its
location controlled the straits of the Gulf of Corinth making it an important entrepot."8 At

35 Leake, op. cit. III p.7.


"7 Idem. Voyage V (1826) 277.
36 F.C. Pouqueville, Travels in the Morea, Albania and Other 38 Idem. Travels 462-3.
Parts of the Ottoman Empire London (1813) I 8.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 351

a second level are towns of I,ooo to 1,2oo houses including Argos, Mistra and Pi
of these three were administered by a Bey and were centres controlling f
productive agricultural plains. The remainder of the nodes on Gennadius' m
Peloponnese are a mixture of settlements and rural crossroads. Fourteen of th
canton capitals administered either by a Bey or Hodja Bashi and were later
centres of the first Eparchies introduced after Greek independence.39 Of the re
should be noted that at least two of the rural cross-roads were served by Khans
The actual result of the SPM analysis is similar to that expected; Tripolis, K
Kalavrita, Leondari, Patras and the node near Megalopolis are the six most con
while an axis of high connectivity extends from Patras to Tripolis and from
Leondari and Karitaina. At a second level of connectivity, this corridor forms a
pattern with legs extending towards Sparta, Androussa, and Elis with a vestig
towards Argo-corinthia.
The analysis of the 1822 road system (Fig. 6.1, Table 6.2) alone accurately pr
primacy of Tripolis and Patras within the network. A zone of high connectivit
the central Peloponnese including within it Tripolis (rank ist of 33), Karitaina,
Leondari and the node near Megalopolis. Similarly, at a second level of con
Gastouni and Pirgos (one of three towns in the second level of size with I,ooo
exert their own connective 'pull', while Mistra (like Pirgos, another town w
houses) also ranks highly (8th of 33).

High (1139-1149) ,
High (90-106) .
Moderate (107-120)
1 Moderate (1174- 1183) 2

FIG. 6. Connectivity diagram of the Peloponnese in 1822. I. Analysis of the road system alone, 2. Connectivity of the
network measured in time.

39 Expedition Scientifique de la More, Paris I834.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
352 G.D.R. SANDERS and I.K. WHITBREAD

DISTANCE MOREA MOREA + SEA SIZE


LINKS
I : /n-1 Rank I 1/n-1 Rank I r/n-1 Rank

1 MEGARA - - - 197 6.15 33 184 5.75 33


2 CORINTH * 37656 1176.75 16 166 5.18 31 153 4.78 32 300 Houses
3 AG. VASILEOS 37506 1172.06 13 152 4.75 28 140 4.37 29 15 Families
4 ARGOS * 37184 1162.00 7 131 4.09 18 119 3.71 19 1200 Houses
5 KOUTSI 37364 1167.62 10 136 4.25 21= 115 3.59 16=
6 KHAN 37124 1160.12 6 110 3.43 9= 103 3.21 7
7 TRIPOLIS ** 36480 1140.00 2 90 2.81 1 87 2.71 1 2500 Houses
8 SPARTA ** 37674 1177.31 18 109 3.40 8 102 3.18 6 1000 Houses
9 MONEMVASIA * 40674 1271.06 31 140 4.37 25 110 3.43 13 350 Houses
10 LEONDARI * 36432 1138.50 1 101 3.15 4 100 3.12 5 290 Houses
11 KHAN 36768 1149.00 5 110 3.43 9= 108 3.37 10= -
12 KALAMATA * 37376 1168.00 11 136 4.25 21= 133 4.15 27 400 Houses
13 NISI 37445 1170.15 12 146 4.56 27 139 4.34 28
14 ANDROUSSA * 37313 1166.03 9 120 3.75 14 115 3.59 16= 300 Houses
15 NAVARINO * 37811 1181.59 20 138 4.31 24 111 3.46 14 300 Houses
16 MODON * 38147 1192.09 27 165 5.15 30 122 3.81 21 440 Houses
17 CORONE * 38663 1208.21 30 174 5.43 32 144 4.5 30 330 Houses
18 ARCADIA * 38155 1192.34 28 126 3.93 15 124 3.87 22 600 Houses
19 PIRGOS * 37572 1174.12 14 114 3.56 12= 115 3.59 16= 1000 Families
20 GASTOUNI * 37712 1178.50 19 113 3.53 11 112 3.5 15 500 Houses
21 CLARENZA 38162 1192.56 29 144 4.50 26 107 3.34 9
22 PATRAS * 37994 1187.31 23 104 3.25 5 97 3.03 4 2500
Familes+/-
23 KASTRON 38114 1191.06 25 133 4.15 19= 126 3.93 24
24 KHAN 38123 1191.34 26 127 3.96 16 120 3.75 20
25 VOSTITSA * 38114 1191.06 25 133 4.15 19= 125 3.90 23 330 Houses
26 METOCHION 37933 1185.40 22 114 3.56 12= 108 3.37 10=
27 KALAVRITA * 37614 1175.43 15 95 2.96 3 89 2.78 2 600 Houses
28 XYLOKASTRON 38064 1189.50 24 136 4.25 21= 130 4.06 26
29 VASILIKATA 37844 1182.62 21 156 4.87 29 147 4.59 31 60 Familes
30 LALLA * 37663 1176.96 17 128 4.00 17 127 3.96 25
31 XROADS 37269 1164.65 8 108 3.37 7 109 3.40 12
32 KARITAINA * 36516 1141.12 4 94 2.93 2 93 2.90 3 600 Families
33 MEGALOPOLIS 36486 1140.18 3 106 3.31 6 105 3.28 8

TABLE 6. Row sums (X), mean (I/n-I) and ranking of individual nodes base
and sea links respectively. The node size, given in estimated number of ho
Leake Travels in the Morea and F.-C. Pouquville, Voyage de la Grice V. * denot
of two cantons.

Although the Peutinger Table analyses the distance results replicated the road
connectivity test, the 1822 map pattern, using time as a weighting factor, is radically
different (Fig. 6.2, Table 6.i). Again the central Peloponnese nodes are the most central,
reinforcing the impression of the primacy of Tripolis (ranked 2nd marginally behind
Leondari). In contrast to the first analysis with its dominant north-south axis, the time
travel accessibility axis runs diagonally from Argos to Nisi and whereas Patras, Pirgos and
Sparta were highly ranked, they are in terms of time relatively remote (ranking 23rd, I4th,

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 347

TRAVELIMN182 Time=Mnuts/Io

0154723968 0154CORINTH2
2345678910 150368247 AG.VSILEO 4530612987 ARGOS4 51360284971KOUTSI5 5136209847 KHAN 1086529347 TRIPOLS 170523964 815362SPART 1392408567 1239LEONDARI0 1604529738 KHAN1 1958064237 198645KALMT2 1978254630 179ANDROUS4 1570269483 5127049PATRS 1496208573 19602KASTRON3 1435892607 15398KHAN24 1052684937 2108VOSTIA5
9015346287 METOCHIN26
1354728690 13584KALVRIT27
4358912076XYLOKASTRN 1836925740 16VASILKT9
1538042697 15082KARITN3 1483097625 1834960725MEGALOPIS
270531964 82MONEVASI9 2018654397 NIS1 2718643059NAVRIO1 2394108576 9MODN1 2537016948 251CORNE7 2806534197 ARCDI 2540169378 PIRGOS 2540319768 253GASTOUNI0 2405619378 42CLARENZ1 2301589764 03LA 2031856479 0XROADS31 TABLE7.Ousethabl,findmropcg tracedownhlumfqis.Tgvxp/I Example:Trios(#7)tNvn1526,.0uwkgchdy distanbyl.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
354 G.D.R. SANDERS and I.K. WHITBREAD

High (87-102)

Moderate (103 -112)

FIG. 7. Connectivity diagram of the Peloponnese in 1822. Analysis of road and sea links.

and I8th respectively). On the other hand, Argos, the third and largest of the
side-ranked city, is 'close' (ranked 7th of 33).
Factoring in sea links created a third pattern still (Fig. 7, Table 6.3). The highe
nodes are almost identical to those for roads alone and the time factored test but with
Mistra replacing the cross-roads near Megalopolis. At the second level of connectivity,
Pirgos no longer exerts its 'pull', rather Monemvasia's gravity extends the stron
north-south axis to the Aegean coast. High values are also recorded for Navarino, Kyllen
and Gastouni reflecting the two former towns' commercial attraction.
Connectivity analysis of Gennadius' 1822 map for the Peloponnese suggests that road
connections and time travel were critical for the size and prosperity of early nineteenth
century towns while sea lanes played a much less important role than in the Roma
period. The time travel results are interesting in reviewing Ibrahim Pasha's campaign o
1825/26. The Egyptian forces landed in Modon in February 1825 and moved to besiege
Navarino and Pylos in March. Seven days passed before the Greek forces marched from
Nauplion, the Greek capital, to relieve the siege and did not arrive long before their defe
a month later. After securing an excellent harbour for his fleet, Ibrahim advanced to th
town of Arcadia and thence to Nisi and Kalamata in early June. The four months spen
taking the southwest Peloponnese were followed by 14 days during which the Egyptia
thrust north-east, by-passing Kolokotronis at Makriplai (I6th June), and appeare
opposite Nauplion at Lerna on I4th June. For the remainder of the campaigning season
Ibrahim based his operations at Tripolis sending units into Lakonia. In I826 Ibrahim

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 355

again used Tripolis as a headquarters and proceeded to waste Achaia, E


Messenia and Lakonia.40 The choice of beach head, the speed with which Ibr
out of Messenia and his choice of Tripolis all seem to be explainable in t
network analyses.
G.D.R. SANDERS AND ILK. WHITBREAD

APPENDIX: COMPUTER PROGRAMS

io REM Matrix Connectivity and Shortest Path Program


20 REM Guy Sanders and Ian Whitbread
30 REM Annual of the British School at Athens
40 RE~M
50 REM - INFINITY=32767 -
60 REM MAIN PROGRAM
70 REM INITPROG
80 GOSUB 280
go REM MATRIXTYPE
'oo GOSUB 470
I o REM MATRIXSOURCE
120 GOSUB 550
130 REM SAVEMODE
140 GOSUB I410
150 REM CALCULATE
160 GOSUB 1510o
170 REM SAVESHORTPATH
180 GOSUB 2000
190 REM RANKSHORTPATH
200 GOSUB 2250
2Io REM FINISH
220 GOSUB 2680
230 SYSTEM
240 REM~
250 REM SUBROUTINES
260 REM INITIALISATION AND SETTING OF OPTIONS
270 REM - INITIALISATION OF PROGRAM -
280 REM INITPROG:
290 OPTION BASE I
300 INFINITY% =32767 : FILESAVE%=o
310 o DEFINT M
320 DEFDBL S
330 REM ERFIL
340 REM ON ERROR GOTO 460
350 REM - INTRODUCTORY SCREEN
360 CLS
370 LOCATE 2,26 : PRINT "Connectivity Matrix Program" : LOCATE 3,38 : PRINT "by"
380 LOCATE 4,26: PRINT "Guy Sanders & Ian Whitbread"
390 BT%= I : BL%=23 : BB%=5 : BR%=55 : GOSUB 2780
400 LOCATE 7,24: PRINT " INFINITY = 32767
4zo BT%=6 : BL%=22 : BB%=8 : BR%=56 : GOSUB 2780"

40 G. Finlay, A History of Greece VI 356-70, 398-400.

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
356 G.D.R. SANDERS and I.K. WHITBREAD

420 RETURN
430 REM - ERROR HANDLING
440 REM ERFIL:
450 REM RUNERR=53 THEN RESUME NEXT ELSE CLS : PRINT "ERROR"; ERR: STOP
460 REM - SELECT MATRIX TYPE: BINARY OR WEIGHTED -
470 REM MATRIXTYPE:
480 BT%=2I : BL%=3 : BB%=24 : BR%=77 : GOSUB 2780
490 LOCATE 22,4: PRINT "Select matrix: binary (.bmx) (I) weighted (.wmx) (2) - I/2"
500 LOCATE 23,4: INPUT BIWT%
51o IF BIWT%<>I and BIWT%<>2 THEN LOCATE 23,4: INPUT BIWT%
520 IF BIWT%=I THEN FILEEXTI$=".B" ELSE FILEEXTI$=".W"
530 RETURN
540 REM - SELECT MATRIX SOURCE: KEYBOARD OR FILE
550 REM MATRIXSOURCE:
560 BT%=2I : BL%=3 : BB%=24 : BR%=77 : GOSUB 2780
570 GAP%=73 : ROWTOP%=22: ROWBOTTOM%=23 : COLUMN%=4 : GOSUB 2890
580 LOCATE 22,4: PRINT "Enter matrix: from the keyboard (I) from a file (2) - 1/2"
590 LOCATE 23,4: INPUT DATASOURCE%
600oo IF DATASOURCE%<> I AND DATASOURCE%<> 2 THEN LOCATE 23,4 : INPUT
DATASOURCE%

610io GAP%=73 : ROWTOP%=22: ROWBOTTOM%=23 : COLUMN%=4: GO


620 FILEEXT$= FILEEXTI$+"MX"
630 ON DATASOURCE% GOSUB 660, o1060
640 RETURN
650 REM - KEYBOARD INPUT OF MATRIX VALUES
660 REM KEYBOARD:
670 MATRIXCHARLEN% =o
680 LOCATE 22,4 : PRINT "Enter the size of the matrix - i.e. the number of row
690 LOCATE 23,4: INPUT MATRIXSIZE%
700oo IF MATRIXSIZE%<2 and MATRIXSIZE%>INFINITY% THEN LOCATE 23,4 : INPUT
MATRIXSIZE%

710o GAP%=73 : ROWTOP%=22 : ROWBOTTOM%=23 : COLUMN%=4: GOS


720 REM DIMENSION
730 GOSUB 1370
740 LOCATE 7,24: PRINT " ENTER MATRIX VALUES "
750 BT%=6 : BL%=22 : BB%=8 : BR%=56 : GOSUB 2780
760 FOR ROW% = i TO MATRIXSIZE%
770 FOR COL% = i TO MATRIXSIZE%
780 LOCATE 22,4: PRINT "Enter value (<32767 : O= INFINITY) for ROW:COLUMN - "; ROW%
;":";COL%
790 LOCATE 23,4: INPUT MATRIXSHORTPATH(ROW%,COL%)
800oo IF BIWT%= I AND MATRIXSHORTPATH(ROW%,COL%)>I THEN LOCATE 23,4: INPUT
MATRIXSHORTPATH(ROW%,COL%)
8io IF MATRIXSHORTPATH(ROW%,COL%)>=INFINITY% THEN LOCATE 23,4 : INPUT
MATRIXSHORTPATH(ROW%,COL%)
820 REM - DETERMINE LENGTH OF MAX. INTEGER TO FORMAT OUTPUT
830 IF LEN(STR$(MATRIXSHORTPATH(ROW%,COL%)))>MATRIXCHARLEN% A
CHARLEN%<6 THEN MATRIXCHARLEN%=LEN(STR$(MATRIXSHORTPATH(ROW%,
COL%)))
840 GAP%=73 : ROWTOP%=22 : ROWBOTTOM%=23 : COLUMN%=4 : GOSUB 2890
850 NEXT COL%
860 NEXT ROW%
870 LOCATE 22,4 : PRINT "Matrix file name: [d:][path]file - file=8 chars max. - n
880 LOCATE 23,4: INPUT MATRIX$
890 GAP% =73 : ROWTOP%=22 : ROWBOTTOM%=23 : COLUMN%=4 : GOSUB 2890
900 REM SAVETODISK

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 357

9IO0 GOSUB 2080


920 LOCATE 22,4: PRINT "Select: Quit the program now (I) Calculate the matrix (2) - 1/2"
930 LOCATE 23,4: INPUT GETOUT%
940 IF GETOUT%<> 1 AND GETOUT%<>2 THEN LOCATE 23,4: INPUT GETOUT%
950 GAP%=73 : ROWTOP%=22 : ROWBOTTOM%=23 : COLUMN%=4: GOSUB 2890
960 IF GETOUT%= i THEN GOSUB 2600
970 GAP%=78: ROWTOP%=6: ROWBOTTOM%=8 : COLUMN%=I : GOSUB 2890
980 FOR ROW% = i TO MATRIXSIZE%
990 FOR COLUMN% = I TO MATRIXSIZE%
100ooo REM - off axis 0 values=infinity
Iolo IF MATRIXSHORTPATH(ROW%,COLUMN%)=0 AND ROW%<>COLUMN% THEN
MATRIXSHORTPATH(ROW% ,COLUMN%) = INFINITY%
1020 NEXT COLUMN%
1030 NEXT ROW%
o1040 RETURN
1050 REM - FILE INPUT OF MATRIX VALUES
Io60 REM FILE:
Io70o GAP%=78:ROWTOP%=6: ROWBOTTOM%=8 : COLUMN%=I :GOSUB 2
lo80 LOCATE 7,I : PRINT "MATRIX FILES AVAILABLE IN CURRENT DIREC
Io90 IF BIWT%=I THEN FILES "*.bmx" ELSE FILES "*.wmx"
I100 LOCATE 22,4 : PRINT "Matrix file name: [d:][path]file - file=8 chars max. - no e
1I10 LOCATE 23,4: INPUT MATRIX$
1120 GAP%=73: ROWTOP%=22 : ROWBOTTOM%=23: COLUMN%=4: GOSUB
1130 LOCATE 22,4 : PRINT "File "+MATRIX$+FILEEXT$+' is being loaded."
1140 OPEN MATRIX$+FILEEXT$ FOR INPUT AS #I
I150 REM - FIND NUMBER OF ROWS TO DETERMINE MATRIX SIZE
MATRIX ARRAYS
1160 MATRIXSIZE% o
I170 WHILE NOT EOF(I)
1180 LINE INPUT#I,CHAR$
1190 MATRIXSIZE%= MATRIXSIZE% +
1200 WEND
12io REM DIMENSION
1220 GOSUB 1370
1230 CLOSE
1240 REM - INPUT AND INITIALISE MATRIX ARRAYS
1250 OPEN MATRIX$+FILEEXT$ FOR INPUT AS #1
1260 FOR ROW% = I TO MATRIXSIZE%
1270 FOR COLUMN% = I TO MATRIXSIZE%
1280 INPUT #I, MATRIXSHORTPATH(ROW%,COLUMN%)
1290 REM - off axis o values=infinity
1300 IF MATRIXSHORTPATH(ROW%,COLUMN%)=o AND ROW%<>COLUMN% THEN
MATRIXSHORTPATH(ROW%,COLUMN%)=INFINITY%
131o NEXT COLUMN%
1320 NEXT ROW%
1330 CLOSE
1340 GAP%=73 : ROWTOP%=22 : ROWBOTTOM%=23 : COLUMN%=-4 : GOSUB 2890
1350 RETURN
1360 REM - DIMENSION MATRIX ARRAYS
1370 REM DIMENSION:
1380 DIM MATRIXSHORTPATH(MATRIXSIZE%,MATRIXSIZE%), SUMROWSHORTPATH-
(MATRIXSIZE%,2)
1390 RETURN
1400 REM - SELECT MATRIX SAVE MODE: AUTOMATIC OR MINIMUM -
1410 REM SAVEMODE:
1420 LOCATE 22,4: PRINT "File save: Automatic (i) Minimum (2)- I/2"

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
358 G.D.R. SANDERS and I.K. WHITBREAD

1430 LOCATE 23,4: INPUT FILESAVE%


1440 IF FILESAVE%<>I AND FILESAVE%<>2 THEN LOCATE 23,4: INPUT FILESAVE%
1450 GAP%=73 : ROWTOP%=22 : ROWBOTTOM%=23 : COLUMN%=4 GOSUB 2890
1460 GAP%=74: ROWTOP%=6 : ROWBOTTOM%=20 : COLUMN%=I : GOSUB 2890
1470 RETURN
1480 REM
1490 REM CALCULATION OF BINARY (o/I) AND WEIGHTED (N=o/INFINITY) MATRICES

1500 REM MATRIX CALCULATION


1510 REM CALCULATE:
1520 LOCATE 7,26: PRINT "MATRIX IS BEING CALCULATED"
1530 BT%=6:BL%=22 :BB%=8: BR%=56 : GOSUB 2780
1540 BT%=9: BL%= Io : BB%= 16 : BR%= 7o : GOSUB 2780
1550 LOCATE 22,4 : PRINT "Matrix = "+MATRIX$+FILEEXT$
1560 LOCATE 23,4: PRINT "Matrix dimensions =";MATRIXSIZE%;"x"; MATRIXSIZE%
1570 REM CALCULATE WEIGHTED MATRIX SHORT PATH FOR EACH NODE
1580 REM - RAISE MATRIX THROUGH SUCCESSIVE POWERS
I590 NODE%= I
1600 WHILE NODE%<>MATRIXSIZE%
I6Io NODE% =NODE%+I : INFINITYSUM%=o
1620 REM ADDWEIGHTED
1630 GOSUB 1700
I640 REM WEIGHTEDINFOSCREEN
1650 GOSUB I9Oo
1660 WEND
1670 FILEEXT2$="SP"
1680 RETURN
1690 REM - WEIGHTED MATRIX ADDITION
I700 REM ADDWEIGHTED:
1710 MATRIXCHARLEN%=o
1720 FOR A% = I TO MATRIXSIZE%
1730 FOR B% = I TO MATRIXSIZE%
1740 IF A%<>B% AND B%<>NODE% AND NODE%<>A% THEN GOSUB 1820
1750 IF A%<>B% AND B%<>NODE% AND NODE%<>A%AND
MATRIXSHORTPATH(A%,B%)= INFINITY% THEN INFINITYSUM%= INFINITYSUM% + I
1760 REM - DETERMINE LENGTH OF MAX. INTEGER TO FORMAT OUTPUT
i770 IF LEN(STR$(MATRIXSHORTPATH(A%,B%)))>MATRIXCHARLEN% AND MATRIX-
CHARLEN%<6 THEN MATRIXCHARLEN%= LEN(STR$(MATRIXSHORTPATH(A%,B%)))
1780 NEXTB%
1790 NEXTA%
i8oo RETURN
I8Io REM - WEIGHTED MATRIX: DETERMINE MINIMUM PATH: CHECK
INFINITY
1820 REM CALCMIN:
1830 MATRIXTEST%=o
1840 IF MATRIXSHORTPATH(A%,NODE%)=INFINITY% OR MATRIXSHORTPATH(NODE-
%,B%) )= INFINITY% THEN MATRIXTEST% = INFINITY%
1850 IF INFINITY% - MATRIXSHORTPATH(A%,NODE%), MATRIXSHORTPATH (NODE%,B%
THEN MATRIX TEST%= INFINITY%
1860 IF MATRIXTEST%<>INFINITY% THEN MATRIXTEST%=MATRIXSHORTPATH(A%,
NODE%) +MATRIXSHORTPATH(NODE%,B%)
1870 IF MATRIXSHORTPATH(A%,B%)>MATRIXTEST% THEN MATRIXSHORTPATH(A%,B-
%)= MATRIXTEST%
1880 RETURN
1890 REM - DISPLAY CURRENT INFORMATION ON WEIGHTED MATRIX CALCULATION
1900oo REM WEIGHTEDINFOSCREEN:

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 359

I9Io GAP%=56:ROWTOP%=I2: ROWBOTTOM%=I4: COLUMN%=II :GOSUB 2890


1920 LOCATE 10,I : PRINT "CURRENT NODE = ";NODE%
1930 LOCATE II,II: PRINT "Number of infinity values =";INFINITYSUM%
1940 REM - SAVE MATRIX FOR EACH NODE IF AUTOSAVE SELECTED
1950 IF FILESAVE%= I THEN FILEEXT$="."+MID$(STR$(NODE%),2) : GOSUB 2080
1960 IF FILESAVE%=I THEN LOCATE 12,11 : PRINT "Node matrix saved - see file "+MATRIX
$+FILEEXT$
1970 RETURN
1980 REM
1990 REM MATRIX FILE OUTPUT SUBROUTINES
2000 REM SAVESHORTPATH:
20oio FILEEXT$=FILEEXTI$+"SP"
2020 REM SAVETODISK
2030 GOSUB 2080
2040 LOCATE I13,I : PRINT "Short path matrix saved - see file "+MATRIX$+FILEEXT$
2050 RETURN
2060 REM - MAIN SAVE SUBROUTINE -
2070 REM - SAVE MATRIX FILES
2080 REM SAVETODISK:
2090 REM - DETERMINE MAXIMUM NUMBER LENGTH IN MATRIX TO FORMAT OUTPUT
2Ioo00 FORMATS$="#"
21i10 FOR A%= I TO MATRIXCHARLEN%
2120 FORMAT$= "#" +FORMAT$
2130 NEXTA%
2140 OPEN MATRIX$+FILEEXT$ FOR OUTPUT AS #i
2150 FOR ROW% = i TO MATRIXSIZE%
2160 FOR COLUMN% = I TO MATRIXSIZE%
2170o PRINT #i, USING FORMAT$; MATRIXSHORTPATH(ROW%,COLUMN%);
2180 NEXT COLUMN%
2190 PRINT #I,
2200 NEXT ROW%
221O CLOSE
2220 RETURN
2230 REM
2240 REM RANK THE SHORTPATH MATRIX ROWS: FROM MINIMUM SUM (= HIGHEST
RANK = i) TO MAXIMUM - SAVE RANKING
2250 REM RANKSHORTPATH:
2260 REM - SUM ROWS IN MATRIX TO GIVE RANKING
2270 FOR ROW%=I TO MATRIXSIZE%
2280 FOR COLUMN%= TO MATRIXSIZE%
2290 SUMROWSHORTPATH(ROW%,I)= SUMROWSHORTPATH
(ROW%, I) +MATRIXSHORTPATH(ROW%,COLUMN%)
2300 NEXT COLUMN%
23o10 NEXT ROW%
2320 REM - PERFORM RANKING
2330 RANK=o: RANKMAX=SUMROWSHORTPATH(I,I) : LASTRANK=o
2340 WHILE LASTRANK<>RANKMAX
2350 RANK=RANK+I
2360 RANKMIN=RANKMAX
2370 FOR ROW%= i TO MATRIXSIZE%
2380 IF SUMROWSHORTPATH(ROW%, I)<RANKMIN AND SUMROWSHORTPATH(ROW-
%,I)>LASTRANK THEN RANKMIN=SUMROWSHORTPATH(ROW%,I)
2390 IF SUMROWSHORTPATH(ROW%, )>RANKMAX THEN RANKMAX= SUMROWSHORT-
PATH(ROW%, I)
2400 NEXT ROW%
24IO FOR ROW%=I TO MATRIXSIZE%

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
360 G.D.R. SANDERS and I.K. WHITBREAD

2420 IF SUMROWSHORTPATH(ROW%,I)=RANK MIN THEN SUMROWSH


)=RANK
2430 NEXT ROW%
2440 LASTRANK=RANKMIN
2450 WEND
2460 REM - FILE OUTPUT OF RANK DATA
2470 OPEN MATRIX$=FILEEXTI$+"RN" FOR OUTPUT AS #i
2480 PRINT # I, "MATRIX =";MATRIX$+FILEEXT$
2490 PRINT #i, "NUMBER OF NODES =";MATRIXSIZE%
2500 PRINT #I, " "
25Io PRINT #i, "ROW" TAB(Io) "SUM OF ROW" TAB (30) "MEAN" TAB(55) "RANK (ORDINAL)"
2520 FOR ROW% = i TO MATRIXSIZE%
2530 PRINT #I,ROW% TAB(Io) SUMROWSHORTPATH(ROW%,I);
2540 PRINT #i ,TAB(3o) SUMROWSHORTPATH(ROW%,I)/(MATRIXSIZE%-I);
2550 PRINT #I, TAB(55) SUMROWSHORTPATH(ROW%,2)
2560 NEXT ROW%
2570 CLOSE
2580 LOCATE 15,11 : PRINT "Row sums, ranks and means saved - see file "+MATRI
$+FILEEXTI$+"RN"
2590 RETURN
2600 REM LEAVE:
26o10 GAP%=74: ROWTOP%=6 ROWBOTTOM%=2o0: COLUMN%=I :GOSUB
2620 REM FINNMESSCLR
2630 GOSUB 2960
2640 END
2650 RETURN
2660 REM
2670 REM COMPLETION OF MATRIX CALCULATION SCREEN
2680 REM FINISH:

2690 BT%=17 : BL%=23: BB%=20o: BR%=56: GOSUB 2780


2700 LOCATE 18,25 : PRINT "All nodes are now connected."
2710 LOCATE 19,25: PRINT "MATRIX COMPUTATION COMPLETED."
2720 REM FINNMESSCLR
2730 GOSUB 2960
2740 RETURN
2750 REM
2760 REM _DISPLAY SUBROUTINES -
2770 REM - DRAW BOX ON SCREEN
2780 REM BOX:
2790 FOR ROW%=BL%+I TO BR%-I
2800 LOCATE BT%,ROW% : PRINT CHR$(205); : LOCATE BB%,ROW% : PRINT CHR$(205);
2810 NEXT ROW%
2820 FOR COLUMN%=BT%+ 1 TO BB%-I
2830 LOCATE COLUMN%,BL% : PRINT CHR$(186);: LOCATE COLUMN%,BR% : P
2840 NEXT COLUMN%
2850 LOCATE BT%,BL% : PRINT CHR$(20I); : LOCATE BB%,BL% :PRINT CHR$(200);
2860 LOCATE BT%,BR% : PRINT CHR$(I87); : LOCATE BB%,BR% : PRINT CHR$(I88);
2870 RETURN
2880 REM CLEAR SELECTED PORTIONS OF SCREEN
2890 REM MESSCLR:
2900 LINECLEAR$=SPACE$(GAP%)
2910 FOR R%= ROWTOP% TO ROWBOTTOM%
2920 LOCATE R%,COLUMN% : PRINT LINECLEAR$;
2930 NEXT R%
2940 RETURN

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms
CENTRAL PLACES AND MAJOR ROADS IN THE PELOPONNESE 36

2950 REM - CLEAR SELECTED PORTIONS OF SCREEN FOR FINAL MESSAGE ON TERMI-
NATION

2960 REM FINNMESSCLR:


2970 LINECLEAR$= SPACE$(32)
2980 LOCATE 7,24: PRINT LINECLEAR$;
2990 GAP%=78: ROWTOP%=7: ROWBOTTOM%=8 : COLUMN%=I : GOSUB
3000 BT%=6: BL%=Io: BB%=8 : BR%=7o : GOSUB 2780
30oo LOCATE 7,12 : PRINT "For files enter: DIR "+MATRIX$+".*"
3020 GAP%=78: ROWTOP%=2I : ROWBOTTOM%=24: COLUMN%=I : GOSU
3030 RETURN
3040 REM

This content downloaded from 83.212.248.205 on Tue, 05 Dec 2017 20:28:24 UTC
All use subject to http://about.jstor.org/terms

You might also like