The reader is advised that as a consequence of this nomenclature confusion, the troctolites described in Unit 6 (Dick, 1991a) around 404 to 550 meters below seafloor (mbsf) in Hole 735B are quite distinct from the troctolites found lower in the hole. The former are all small meter-scale intrusions with sharp crosscutting contacts with the olivine gabbros that grade from microtroctolite (fine grained) to troctolite (medium to coarse grained) under this classification. They are the most primitive rocks recovered from Hole 735B—apparently having crystallized from a different or at least more primitive parental magma than those giving rise to the olivine gabbros. By contrast, the "troctolites" deeper in the hole simply represent a continuum from gabbro, olivine gabbro, and troctolitic gabbro, representing an increase in the proportion of modal olivine. Otherwise, these latter rocks are texturally and mineralogically identical.

Medium- to coarse-grained leucocratic igneous rocks make up ~1% of the Hole 735B core. They are generally present as crosscutting small veins and are commonly very altered. In many cases, they are hard to distinguish from purely hydrothermal "felsic" veins. Although they range from diorite through tonalite to trondhjemite and include a few granite veins, the specific igneous protolith is often very difficult to identify. Thus, many specific identifications made by the analysts and the shipboard party must be regarded as somewhat suspect. Many readers will be more familiar with these rocks as "plagiogranites," a common field term for white felsic igneous rocks of ambiguous nature, specifically not recommended by IUGS as it overlaps with other long-used valid rock names. Because we would prefer to leave the final identification of most of these to those with some specialization in the field of late igneous leucocratic magmatic veins, we have lumped them in the tables collectively under the name "felsic vein." There is no denying the usefulness of the term plagiogranite in this context when trying to deal with these rocks, but because there is little agreement among ourselves over the term "plagiogranite," we do not use it here.

The complexity of the terminology and diversity of rock names used will undoubtedly intimidate the reader—as it does the authors. Whereas good reasons exist for each name and most turn out to have genetic significance, we often simplify this terminology in this paper. The reader is advised that when we use the plural form we are generally referring to a clan of rocks with petrologic affinity. By gabbros we mean collectively all the gabbroic rocks in the hole, excluding the crosscutting microgabbros. When we refer to olivine gabbros we mean collectively all the oxide-poor olivine-bearing gabbros (gabbro, olivine gabbro, and troctolitic gabbro). Similarly, oxide gabbros, or the term "ferrogabbro," refers to the many varieties of oxide-rich, orthopyroxene-poor olivine-bearing and olivine-free gabbro. Gabbronorites include all the rock types with significant amounts of granular orthopyroxene (gabbronorite, orthopyroxene gabbro, orthopyroxene-olivine gabbro, pigeonitic gabbro, etc.). Specific rock types are generally referred to in the singular. Troctolite and disseminated oxide olivine gabbro are unique in this respect, as they are generally not included in the three general rock clans (olivine gabbros, gabbronorites, and oxide gabbros). Many of the troctolites are distinct from all other rock types in various compositional plots, and the disseminated oxide olivine gabbros appear to be transitional from the olivine gabbro clan to the oxide gabbro clan without having significant orthopyroxene.

Columns N through S give the average Mg# (molar Mg x 100/[Mg + Fe*]) and forsterite, anorthite, and orthoclase contents of coexisting olivine, clinopyroxene, orthopyroxene, plagioclase, and brown hornblende in the thin section analyzed, except for the analytical point whose measured values are given. Modal data are given in columns T through AC, where available. Data for the ODP thin sections were collected by Dr. Peter Meyer during Leg 176 and are taken from the Leg 176 Initial Reports volume. These data were used to assign rock names to all the ODP thin sections analyzed by H. Dick using the criteria laid out in the Leg 176 Initial Reports volume. Modes for the Y. Niu samples were collected by S. Mackie, whereas modes given for the R. Hebert and M. Constantin samples are visual estimates provided by them. Modes for the J. Maeda samples were also estimated by H. Dick. "Other" refers largely to metamorphic replacements of primary minerals, as can be seen where the sum of the major igneous minerals is <100. This presents a problem that the user should be aware of insofar as most of the remaining data represent simple "primary modes" where the replacements are lumped with the amount of relict igneous mineral phase. A quantity of 0.1 has been arbitrarily used wherever an analyst gives a modal estimate of "trace."

Textural terms are difficult to quantify, as they are often quite subjective, particularly for igneous rocks. However, column AD gives codes corresponding to the deformation microfabrics and intensity of crystal-plastic deformation intensity seen in many thin sections as well as the presence or absence of an igneous lamination. These are based on the guidelines set forth in table T4 in "Structural Geology" in the "Explanatory Notes" chapter of the Leg 176 Initial Reports volume. These textural codes were assigned by the structural geology team to all the Leg 176 shipboard thin sections, and we have taken their data from the Leg 176 Initial Reports volume. In addition, J. Hirth, who was a member of the Leg 176 structure team, has further described his, W. Bach's, and H. Dick's samples with Dick; and Dick has described the textures of J. Maeda's and most of P. Robinson's samples using the same criteria. In addition, M. Constantin provided unpublished textural data for their samples that permitted H. Dick to estimate the deformation grade, which has been included in the tables.

Textures 1a and 1b refer to a largely undeformed simple igneous texture without and with an igneous lamination, respectively (usually defined by a feldspar-shape fabric). Textures 2, 3, and 4 refer to the intensity of crystal-plastic deformation. Texture 2 represents <30% recrystallized plagioclase with only minor crystal-plastic deformation of pyroxene and relict igneous fabric preserved, whereas texture 3 represents >30% recrystallization of plagioclase with moderate recrystallization of pyroxene and a moderate shape-preferred orientation fabric. Texture 4 represents an extensively recrystallized rock with strongly bimodal grain size distribution with plagioclase, clinopyroxene, and olivine, all extensively recrystallized with bent and kinked porphyroclasts of the original igneous minerals. Textures 5 and 6 refer to the presence of semibrittle and brittle/cataclastic fabrics. The latter can be superimposed on either crystal-plastic or igneous fabrics, and so in some cases there may be two codes given in the CP field.

Detailed information on the analyses are given in columns AE through AI, with "pt" being the analyst's analytical point number, which may start over with each grain or may represent the sequential point number for an entire analytical run on the microprobe. Where the analyst did not give a point number, we have arbitrarily assigned one. Where the row in the table represents an average or the standard deviation of the average rather than a single analytical point, the number given is the number of analytical points in the average. Column AF gives the grain number, prefixed by a two- or three-letter code for the mineral analyzed. This allows the user to combine and sort data from the different tables. Column AG gives the location of the point analyzed on the grain. C = core, R = rim, and i1, i2, i3, and so on = successive analytical points between the core and the rim numbered sequentially outward (thus, "C, i1, i2, i3, R" represents five sequential analyses from core to rim). B = an isolated point somewhere within a grain. In column AG, neo = neoblast, and ANe and sdNe = averages and standard deviations for neoblast analyses. AC, sdC, AR, and sdR = averages and standard deviations for multiple analyses on cores and rims of grains. Whereas standard deviations are given for all sets of two or more analyses on a single grain, the user is advised that a standard deviation for two points is not very meaningful. Pt = a single analytical point on a grain where no additional spots were analyzed, and thus no average or standard deviation is given for the grain. This is often referred to in the text as a "spot analysis." Columns AH and AI give the rough dimensions of the grain analyzed in millimeters or a simple estimate of the grain size: F = fine grained (<1 mm), M = medium grained (1-5 mm), C = coarse grained (5-30 mm), and P = pegmatoidal (>30 mm). Where only a single letter code is given, it generally is for the average grain size of the rock. When two-letter codes are given, the first code is for the grain analyzed and the second for the average grain size of the rock.

Average analyses for individual mineral grains and their standard deviations are given in the tables. Different analysts, however, used different analytical schemes with different goals in mind. For the most part, the averages given in the tables represent a single grain. Several analysts, however, notably P. Robinson, K. Ozawa, and P. Meyer, were generally attempting to obtain average compositions for the thin section rather than averages for individual grains. For P. Robinson and P. Meyer, we have recast their data such that averages represent average compositions for individual grains, though this often produces many averages and single-point analyses, whereas for K. Ozawa, there are insufficient data to identify the individual grains analyzed and so all analytical points for a given mineral from a thin section are averaged in the tables. The user seeking average compositions for thin sections may want to recombine and average the data by thin section. A point of caution, however, is advisable here. The Hole 735B core is highly heterogeneous, with nearly a thousand discrete igneous intervals identified in hand specimen by the shipboard scientific parties. Additional subdivision on the centimeter scale is possible (e.g., Natland, Chap. 11, this volume), and it should be noted that there are abrupt changes in lithology and mineral chemistry present on the centimeter and millimeter scale. Thin sections commonly cross boundaries between different lithologies, so averaging compositions by thin section may be inappropriate for many samples.

The first author sought to systematically document the extent of chemical zoning present in a thin section. For this purpose, grains were selected for analysis representing a combination of coarse grain size and the greatest optically apparent zoning, with usually only a single grain analyzed for each rock type. Thus, for plagioclase, a single grain was usually analyzed at six locations, starting at the core and moving sequentially out to the rim. For pyroxene, where exsolution was often present, a 10-µm spot size was used with 10 adjoining spots analyzed in a traverse across the plane of exsolution in an attempt to obtain the primary igneous composition. Wherever the grain size was suitable, a separate traverse was made on the core and the rim of the pyroxene grain. Thus, most of H. Dick's analyses represent pyroxene core and rim compositions, rather than bulk grain composition. The effect of this can be seen in many of the downhole plots presented later in this paper, where the small offset between core and rim compositions makes them plot together as an oval—giving the impression that the plot is blurred. In many cases, however, particularly with orthopyroxene, the grain was interstitial to olivine, plagioclase, and clinopyroxene and only one traverse was made, representing a bulk composition. Users are advised to use column AE to sort out averages representing a satisfactory number of points on a given grain if they are attempting to evaluate the extent of heterogeneity on a grain scale.