The "Observation" number simply corresponds to the row number of the original data point. That is, you will have 10,000 observations given your data set. Note that the component scores shown in the data tip may not match the scores returned by "pca" according to the "biplot" documentation:
biplot scales the scores so that they fit on the plot: It divides each score by the maximum absolute value of all scores, and multiplies by the maximum coefficient length of coefs. Then biplot changes the sign of score coordinates according to the sign convention for the coefs.
In R2015a, the exact transformation steps can be seen in lines 198 and 199 of "biplot.m":
maxCoefLen = sqrt(max(sum(coefs.^2,2)));
scores = bsxfun(@times, maxCoefLen.*(scores ./ max(abs(scores(:)))), colsign);
