Thanks for all the Euclidian advice.. I’m hoping that this version, with the offending paragraphs replaced by the bold section, will meet the stringent requirements. More fun than anticipated.
Since a cycloid is traced out by a point on the rim of a circle rolling on a straight line, it should be possible to reconstruct that circle and hence a tangent, as follows. Join LN to get the straight line mentioned, and construct the perpendicular bisector of LN crossing LN at M and the cycloid at K. Construct the perpendicular bisector of KM crossing KM at C. This is the path of the centre of the rolling circle. Draw a ‘central circle’ with centre at C and diameter KM. Construct a line from P, perpendicular to KM, to first meet this central circle at a point Q. Draw in the line MQ and construct a line through P perpendicular to MQ. This is the tangent to the cycloid at P.
[N.B. the position of the rolling circle is that of the ‘central circle’ displaced by a distance PQ. MQ is therefore parallel to the former RP where R is the centre of rotation of the rolling circle. RP is no longer needed as part of the construction.]
Incidentally, this plan involves constructing perpendicular lines. I got the feeling that no-one objected to that process in the first offering, but I’d be interested to know how it’s done without lifting the compasses (and therefore losing their setting)?