0° +23.6° 
EQUINOX OR MID — POSITION SUMMER SOLSTICE 
OF THE SOLAR/LUNAR CYCLES 
Solar markings at midday 
  
WINTER SOLSTICE 
-23.6° 
  
   
  
  
  
   
  
  
0° +18.4° 
EQUINOX OR MID — POSITION MINOR STANDSTILL 
OF THE SOLAR/LUNAR CYCLES OF THE MOON 
Solar/lunar markings at rising 
MAJOR STANDSTILL 
OF THE MOON 
  
+28.7° 
  
  
  
Figure 3: The solar and lunar markings at the spiral of the Sundagger-Site. 
Right: Formation of solar and lunar shadowl/light patterns by the three slabs near 
meridian passage (upper) and at rising (lower). Left: Schematic of the resulting patterns 
on the spirals at the indicated declinations and seasons (see Sofaer et al. 1979,). 
Illustration by Pat Kenny, cc: The Solstice Project (Sofaer et al., 1986). 
  
  
  
  
  
  
  
  
  
  
  
Date 1979 1983 1986 1986 1987 
Camera Kelsh K-470 Wild P-32 Hasselblad Rollei Rollei 
Calibration No Yes Bad No Yes 
Targets Yes Yes Yes Yes Yes 
Control No Yes No No No 
Coverage General & Everything Everything Interior General 
Lunar Edge 
Image Quality Good Excellent Good Good Good 
  
  
  
  
Table 1: Comparison of the photographs of the Sundagger-Site captured by different 
parties (courtesy of VEXCEL, Boulder, CO). 
The strict analytical derivation of a triangular network 
of that kind through given reference points is an ambiguous 
problem. It would have required a lot of programming and 
developmental effort. Furthermore, we wanted to take 
advantage of existing software and avoid the 
implementation of new functions. This was especially 
important as we wanted to visualize the surfaces in 
perspective views. 
We selected Intergraph's InRoads program to create the 
triangular network, and Model View to display the shaded 
model. As InRoads is limited to single surfaces above the 
datum plane the points representing the slabs had to be 
separated into various subsets, each based on an individual 
reference plane. 
Once the triangulation of each subset was completed in 
a local reference system, the points were transformed back 
into the original system. The connections between points, 
which actually define the triangular surface, were 
maintained. Figure 4 shows a shaded view of the 
Sundagger Site as it was created by MicroStation. 
To visualize all details of the site a video-animation was 
created, which moves the viewer all around the slabs and 
even in between the gaps so that the spiral can be seen. 
Furthermore, we experimented with moving each slab 
separately. This was used to demonstrate the possible 
creation of the Sundagger Site. The formation of the three 
slabs was originally attached to the overhanging rock to the 
left of the existing site. Archaeologists assume that the 
slabs fell straight down and were somehow moved to the 
current upright position. Again, we created a video- 
animation to demonstrate the theorized motion of the rocks 
in the computer. 
5. SHADOW CASTING 
In order to better visualize and verify the functioning of 
the Sundagger Site, the casting of the shadows that create 
the dagger on the spiral was simulated. This can be done 
by the ModelView software using the 3-dimensional 
surface model generated previously. Due to some 
limitations of this package, such as the requirement to 
define illumination angles to the nearest degree, and the 
very time-consuming ray-tracing which is performed to cast 
shadows, we looked for an alternative. 
A few years ago, when the first attempts were made to 
reconstruct the Sundagger Site, a shadow-casting program 
was written by Erich Brechner, a computer-engineer now 
employed by Boeing. It runs on Silicon Graphics 
workstations and casts the shadows of the rock onto a user 
defined plane. The direction of the illumination source can 
be defined either as a date and time for the sun, or as the 
lunar declination. The surface of the rocks is not needed 
for this program, which finds the outline of the projected 
shadows simply from the group of points that defines a 
slab. The combination of these shadows results in the 
sundagger projected onto the spiral which is lying in the 
projection plane. Although this is a simplification of the 
real situation, as the spiral is not completely flat, but rather 
spherical, the results proved satisfactory.