Deep drill holes bored into in situ oceanic crust were reviewed. Since the beginning of the Deep Sea Drilling Project, 45 drill holes have penetrated over ca. 50 m into normal oceanic crust, however, most are concentrated in the north Atlantic and east Pacific Oceans. Basement ages of most holes are younger than 20 Ma, and are rather biased toward younger crust compared to the average age of the oceanic crust (61 Ma). Only five deep drill holes have penetrated over 500 m into the basement, three of which drilled into slow spread crust formed at <4 cm/yr, with only one hole is fast spread crust formed at >8 cm/yr and one in intermediately spread crust. Three deep holes (332B, 395A, 418A) drilled into slow spread crust formed at Mid-Atlantic Ridge gave the first evidence of magnetic reversals through the vertical oceanic crust, and showed that the slow spread upper oceanic crust away from hot spots is dominantly composed of pillow lavas with a normal MORB-like affinity. Younger 332B and 395A holes (3.5 and 7.3 Ma) gave poor core recovery (18-21%), while the oldest Hole 418A (110 Ma) yielded a fairly high recovery of 72%. Hole 504B is the only hole to penetrate the extrusive rocks and most of the way through the sheeted dike complex (1836.5 m sub-basement). Average core recovery dropped from 29.8-25.3% in the lava and transition zone down to 14.3% in the sheeted dike complex. Unfortunately, the 504B lava is the depleted extremity of MORBs from intermediate-fast spread ridges. One of the most important findings of Hole 504B is a discrepancy between the seismic velocity structure and the downhole lithology in that the Layer 2/3 boundary resides in the middle of the sheeted dikes, as interpreted by the difference in porosity and bulk density. Hole 1256D is dedicated to coring typical oceanic crust and ultimately penetrates the entire crust into the upper mantle. The site is located on the 15-Ma Cocos plate generated at a superfast rate (22 cm/yr). 502-m-long cores of basement (48% recovery) are lavas showing moderately evolved MORB-like compositions similar to those from the present fast spread ridges. The hole has been cleaned and left ready for future drilling, possibly into Layer 3. The above examples of deep drill holes show that the major obstacles to ultradeep drilling are hole collapse and poor core recovery. Riser drilling is expected to overcome these obstacles for "the 21-century Mohole".