ABSTRACT

In his discussion of our paper (Pett and Walker, 1971), Allen raises three problems: first, that the Cloridorme Formation may not be a "representative clastic turbidite unit"; second, that there are serious difficulties in defining grain size and using the data to interpret flute burial and Bourns division formation; and third, that our parameter B (=L_md/D_max, where L_md is the length from nose to point of maximum depth and D_max is maximum depth) is not "essentially the same" as his B (eq. 1 of this reply), originally defined from an infinite negative step as B=L/H, where H is the height of the step and L is the distance from the step to the point of flow reat achment. We will comment on each of these problems in turn.

The bulk of our data comes from the Cloridorme Formation because of its superb exposure relative to all other turbidite formations in North America. Flute measurement in other localities proved so slow and difficult that the original plan of sampling many more formations was abandoned. The Cloridorme Formation is a very thick sequence of Middle Ordovician turbidities, and on casual inspection, appears identical to such classic formations as the Aberystwyth Grits, the Normanskill and Martinsburg Formations of the Appalachians, the Stanley and Jackfork of the Ouachitas, and the Haymond and Tesnus of the Marathon Basin. The outstanding feature of the Cloridorme is its superb continuous exposure on the Gaspe coast for more than 100 miles: as a result, it has been studied in far more detai than any other single turbidite formation (Enos, 1969a, b; Walker, 1969, 1970; Parkash, 1970; Parkash and Middleton, 1970; Skipper, 1971: and theses by Enos, 1965; Parkash, 1969; Bhattacharjee, 1970; Pett, 1970; and Skipper, 1970). The features of the Cloridorme which make Allen suspect that "comparatively ... it may be unrepresentative" are those which require continuous outcrop and intensive study for their detection. We feel that many other classic turbidite formations would show the peculiarities of the Cloridorme if enough detailed work were done: nevertheless, we clearly cannot deny that about 85 percent of our total data comes from only one formation.

The main points of Allen's discussion involve grain size. We failed to point out specifically in our paper that the chips we took came from the base of each bed (not from within the flutes). Thin sections were not cut from the 1511 chips for reasons of expense, and time required for complete size analysis. We believe that the mean of the 10 largest grains gives an approximation to Allen's (1968, p. 12) vaguely defined D, "a diameter characteristic of the coarsest fractions of the sediment load carried along in the flow." His criticism that this parameter measures the largest and rarest grains, which could have been trapped in the flutes, is inapplicable because our samples did not come from within the flutes. Our measurement of size, using a binocular microscope with eyepiece g id, was checked by three operators, and the results were tabulated in a two-way analysis of variance model. There was no statistically significant difference between

End_Page 1156-----------------------

operators, but there was between samples. Unfortunately we did not test for one operator making replicate measurements on the same set of samples at different times.

Allen has suggested that it would be important to repeat our work using a different measure of maximum size, say the coarsest 10-percentile. We agree that this would be interesting, but we do not believe that it would give such a meaningful result. In turbidity currents, where suspension of grains is very important, the beginning of deposition and flute burial will be reflected by the coarsest grains. Also, it is known that in many graded beds, the main variation is in maximum size, rather than the average (Passega, 1957; Middleton, 1962, 1967). For example, Middleton (1967, p. 495) has shown that in high concentration experimental turbidity currents, "segregation of grain sizes is also precluded except for some segregation of the largest grain sizes." Hence we believe that the 1-perc ntile (Passega, 1957) or the mean of the 10 largest grains, gives a better indication of the coarsest grains initially burying the flutes than the 10-percentile. Average or modal estimates, as Allen points out, are not particularly suitable. Consequently, the modal figures quoted by Ricci Lucchi (1969) do not constitute a proper field test of Allen's theory, although Allen states that "the investigation, by Ricci Lucchi, affords evidence in favour of a relationship" (between grain size and Bouma division). On the one hand, Allen is claiming that Ricci Lucchi's modal grain sizes afford evidence in favour of his theory: by implication, therefore, Alien is accepting on the other hand that the modal grain sizes (previously stated to be "not particularly suitable") are estimates of D, he "diameter characteristic of the coarsest fractions." Unfortunately, Ricci Lucchi's data cannot be used in this discussion because we do not know the relationship between the modal class and an estimator of the "diameter characteristic of the coarsest fractions," be it the 1, or 10-percentile, or the mean of the 10 largest grains.

Despite the importance of defining a good measure of the coarsest grains in the deposit, the discussion above is much too idealized. Our original criticism of Allen was based on the fact that he equated Vg, the settling velocity of grains within a flute, with a measure of grain size, D, which could be measured in millimeters ... "... thus [by measuring the coarsest fractions] we are provided with a value for D which can alternatively be expressed as a value for V_g" (Allen, 1968, p. 13). In a case where there is no interaction between grains, this may be so, but in a turbidite, it is presently impossible to estimate settling velocity from grain size because of the effects of high grain concentrations and the effect of suspended fines on the fluid viscosi y. Allen's arguments are derived from a consideration of competence, and ignore capacity. Consequently his discussion emphasizes details, such as the choice between the 10-percentile or mean of the 10 largest grains, and ignores the central issue of whether his equation 5 (Allen, 1968, p. 9: eq. 1 below), containing the term V_g, can ever be applied to rocks to calculate U_crit the critical flow velocity of the external stream for continuance of flute marks:

U_crit = [0.91V_gB]/[0.039B - 0.09]

The third criticism by Allen involves our parameter B (= L_md/D_max). We stated that "the point of reattachment in a flute will be the point of maximum scouring, that is, very close to the point of maximum depth" (Pelt and Walker, 1971, p. 125). As Allen points out, this disagrees with his "schematic pattern" (Allen, 1969, fig. 13), and he emphasizes in his discussion of our paper that "reattachment in flute marks does not occur at the point of maximum depth of the mark, but well downstream from this point." However, in his original paper, Allen stated (1969, p. 618) that "flow separated at the rim of each mark and that reattachment occurred close to [our italics] where the trough of the mark attained its lowest elevation." It is impossible to onduct a field test of Allen's theory if he cannot define less ambiguously where the flow actually does reattach--"close to" or ''well downstream" from the point of maximum depth. We have not yet seen Allen's 1971 paper in Sedimentary Geology, which hopefully will give more details of where the flow reattaches.

Nevertheless, the severe criticism of Allen's equation (eq. 1) is based partly on the values of B, but also on the fact that Allen's B is derived from a consideration of infinite negative steps. Flutes do not resemble negative steps because of the highly curved shape of their nose, and Allen himself (1968, p. 9-10) has commented that "interpreted cautiously ... the results given by equation 5 (eq. 1 of this reply) would be valid for blunt-nosed flutes ... any revised criterion for finely skewed flutes ... must be rather tentative."

Allen concludes his discussion by stating that "the score stands even." We do not believe this to be so, because Ricci Lucchi's modal grain size data is inapplicable. However, the reader is the referee, and as such, should keep a close eye on who scores the next goal. We also suggest that "theoretical studies should be designed with due regard for the empirical aspects of the question" (with apologies to Allen for misquoting him!).

End_Page 1157-----------------------