The first half of this article was published in the February 1997 issue of BENCHMARKS under the title: The Tunnel of Eupalinos, a Revisit. Publishing problems at the time caused part two of the story to fall through the cracks and it was never published. Here it is the article in its entirety with some corrections, additions and deletions.
In the April 1991 issue I wrote about the tunnel of Eupalinos on the Greek island of Samos. By chance I recently  obtained a copy of a paper that was published in Germany in 1984 (* see note). It seems that the Samos tunnel had fired the imagination of the public as well as archaeologists and engineers since it was first discovered in 1881. In 1959 German surveyor Wolfgang Kastenbein triangulated across the hill that was pierced by the tunnel, and from an accurate determination of the position of both tunnel entrances calculated the theoretical tunnel axis. From 1971 to 1974 archaeologists of the German Archaeological Institute (DAI) in Athens, Greece, excavated and cleared the entire tunnel and the connecting aqueduct, and in the following year resurveyed the entire project. The resurvey disclosed that the work of the original surveyor, whether it was Eupalinos personally or somebody he employed, was worse than I had described tongue in cheek in my 1991 article.
No satisfactory explanation has come to my attention as to why Eupalinos chose to drive a tunnel through a hill when an open (or covered) canal could have skirted it on the contour at less than double its length. Nor am I convinced that the risk of tunneling from opposite sides (considering the poor survey kills demonstrated by Eupalinos or his surveyors) was worth the savings in construction time versus digging from the upstream side only.
The tunnel pierces Mt. Kastro 540 ft. below its summit. The cross section is almost uniformly 5.9 ft. by 5.9 ft. and the channel that carried the water is about 24 inches wide. Measured along the centerline as constructed, the tunnel is 3,450.15 ft. long, or almost 68 ft. longer than planned. These 68 extra feet are the result of deviations from the planned axis, and the meandering attempt to get the north and south adits to meet.
Kastenbein determined the theoretical tunnel axis to be 3,382.5 ft. long, extending NNW to SSE. He found the difference in elevation between both entrances to be only 1.87 ft. I like to remind my readers that the sole purpose of the tunnel was to carry water in an open channel (covered to keep out debris but not under pressure) sloping about 0.5%, a minimum for such a project. This represents a leveling error of 15 feet. While the nearly level tunnel may have shortened the construction length by roughly 230 ft. it caused an enormous increase in the excavation for the water channel. Since the tunnel entrance was already at least 10 ft. too high as evidenced by the elevation of the approaching canal, the volume of the material dug out of the portion of the channel that is inside the tunnel (dictated by the need to keep the water flowing) is 7% greater than the entire excavated material coming out of the tunnel itself.
The first 845 ft. of the north adit and the first 1,318 ft. of the south adit are straight. At these points crew "North" was about 35 ft. east, and crew "South" about 50 ft. west of the true axis. That fact would not have been known to the diggers, so why was it necessary at that point to change direction? It is possible that the surveyor had noticed a difference of about two degrees in the bearings of the two adits that were still 1,220 ft. from meeting each other, and he was beginning to get a queasy feeling in the stomach. Neither adit was on course and by continuing on their headings they would have missed each other at the midpoint by about 275 ft. Eupalinos was well on his way of digging two tunnels. The discovery probably created enough of a headache to cause him to discontinue the work in the south adit.
There now commenced a zigzag digging in the north adit. The fact that the heading in the north adit was changed several times seems to indicate that the surveyor was not certain whether he was to the east or to the west of the axis, but he must have known from the size of the hill that he was still a long way from making a connection with the south adit. Already too far east, a small correction was made and for the next 68 ft. the miners dug even further into the wrong direction. After another and more drastic change in bearing, this time to the west, the tunnel crossed the (to the surveyor unknown) true axis, and 440 ft. further down its course it was now about 180 ft. west of where it should have been. Having aimed for zig the surveyor now headed for zag, and in the following 440 ft. crossed the axis for the second time and ended up 100 ft. to the west of it. The last 237 feet were dug in an irregular line that was at first more or less parallel to the axis. At some point along their advance the northern miners may have heard the faint echo of tapping signals from the south adit because suddenly the north adit veers toward it. Digging in the south adit was now resumed; also guided toward the hammering sounds of their opposites, and 101 ft. later they met at last.
As mentioned in my 1991 article, admiration of the Samos tunnel dates back to Herodotus’ time (c.484 to c.425 B.C.) and in some quarters continues to this day. The tunnel has been called the eights wonder of the world, a sobriquet that this writer would have liked to attach to more than one survey he has had the misfortune to encounter.
Why does the tunnel deviate as much as 100 feet from its axis? A valid question indeed. It leads to interesting speculation as to how the tunnel may have been surveyed. Author Peters suggested that the zigzag course was deliberate to assure that the straight advancing south adit would not be missed, but even at the second "zag" the south adit was over 500 ft. too short to make a connection possible. Were the surveyors unable to compute the true length of the tunnel from exterior measurements? Very puzzling, to say the least.
The geometry and surveying instruments of the Greeks were adequate for the task at hand, needing only a surveyor with sufficient competence to employ them. Would Eupalinos have attempted to tunnel from opposite sides of a hill had not the surveying skills and methods of his time given him a reasonable assurance that the two adits would meet? Meet they did of course, but only after torturous meandering through solid limestone, hardly an indication of great competence. Compare that to the 1500 feet long Siloam water tunnel in Jerusalem that was dug around 700 BC out of solid bedrock by two teams of diggers who did meet.
I admire the engineering achievements of the Antique but I have my problem with this Samos tunnel. Someone should have advised Eupalinos to find himself a better surveyor or else to dig it all from the upstream side.
* Note: Konrad Peters: Der Tunnel, das Eupalineion auf der Insel Samos. Dortmund, Germany, 1984. In his article, author Peters does not give sufficient data to allow for a precise determination of the deviation of the tunnel centerline from its theoretical axis. My figures are a result of scaling from the included drawings and are therefore not precise.