Modification of fluid inclusions in quartz by deviatoric stress I: experimentally induced changes in inclusion shapes and microstructures

Fluid inclusions in quartz are known to modify their shapes and microstructures (textures) during weak plastic deformation. However, such changes have not been experimentally demonstrated and criteria are not available to relate them to paleostress conditions. To address these issues, quartz crystals containing natural CO₂–H₂O–NaCl fluid inclusions have been experimentally subjected to compressive deviatoric stresses of 90–250 MPa at 700°C and ~600 MPa confining pressure. Strains of up to 1% cause the inclusions to develop irregular shapes and to generate microcracks in crystallographic planes oriented subperpendicular to the major compression axis, σ ₁. The uniform alignment of the microcracks imparts a planar fabric to the samples. The microcracks heal and form swarms of tiny satellite inclusions. These new inclusions lose H₂O by diffusion, thereby triggering plastic deformation of the surrounding quartz via H₂O-weakening. Consequently, the quartz samples deform plastically only in domains originally rich in inclusions. This study shows that fluid inclusions deformed by deviatoric stresses may indeed record information on paleostress orientations and that they play a key role in facilitating crystal-plastic deformation of quartz.     

Modification of fluid inclusions in quartz by deviatoric stress. III: Influence of principal stresses on inclusion density and orientation

Extraction of useful geochemical, petrologic and structural information from deformed fluid inclusions is still a challenge in rocks displaying moderate plastic strain. In order to better understand the inclusion modifications induced by deviatoric stresses, six deformation experiments were performed with a Griggs piston-cylinder apparatus. Natural NaCl–H₂O inclusions in an oriented quartz crystal were subjected to differential stresses of 250–470 MPa at 700–900 °C and at 700–1,000 MPa confining pressure. Independently of the strain rate and of the crystallographic orientation of the quartz, the inclusions became dismembered and flattened within a crystallographic cleavage plane subperpendicular to σ ₁. The neonate (newly formed) inclusions that result from dismemberment have densities that tend towards equilibrium with P _fluid = σ ₁ at T _shearing. These results permit ambiguities in earlier deformation experiments on CO₂–H₂O–NaCl to be resolved. The results of the two studies converge, indicating that density changes in neonate inclusions are promoted by high differential stresses, long periods at high P and high T, and fluid compositions that maximize quartz solubility. Neonates spawned from large precursor inclusions show greater changes in density that those spawned from small precursors. These findings support the proposal that deformed fluid inclusions can serve as monitors of both the orientation and magnitude of deviatoric stresses during low-strain, ductile deformation of quartz-bearing rocks.     

Synthetic fluid inclusions in natural quartz. II. Application to PVT studies

Synthetic fluid inclusions have been used to determine volumetric properties of H₂O and a 20 wt.% KCl solution from 300–700°C and 1–3 kb. Comparison of results obtained for H₂O with previously published data indicates that the synthetic fluid inclusion technique provides PVT data that are within a few percent of those obtained from conventional PVT studies. This technique thus provides a fast, relatively simple means of determining PVT properties of many fluids of geologic interest.     

Compositional re-equilibration of fluid inclusions in quartz

Compositional modifications to salt-water fluid inclusions in quartz were observed following exposure to disequilibrium conditions in a series of laboratory experiments in which samples containing inclusions of known composition were annealed at 3 kbar and 600≤T≤825°C in the presence of fluids having different compositions for time intervals ranging from a few days to one month. Changes in inclusion compositions following re-equilibration were monitored using salt crystal dissolution temperatures and/or IR (infra red) spectroscopy. The behaviors of both synthetic and natural fluid inclusions were studied. The synthetic samples were re-equilibrated under P _int=P _conf conditions to minimize stress in the crystal surrounding the inclusions, and were subjected to both f _H2O ^int f _H2O ^conf and f _H2O ^int f _H2O ^conf . After re-equilibration for four days at T≥600°C, most inclusions displayed significant compositional changes Without decrepitation. Salinity variations as large as ≈ 25 wt% were inferred for brine inclusions exposed to f _H2O ^int ≠f _H2O ^conf for 16 days at 825°C. The majority of our observations are consistent with the net diffusion of water toward the reservior having the lowest μH₂O; i.e., synthetic NaCl−H₂O fluid inclusions exhibited increased Tm(NaCl)s (implying lower relative H₂O contents) when re-equilibrated in the presence of fluids having lower μH₂O, whereas, similar (and, in some cases, the same) inclusions exhibited decreased Tm(NaCl)s (implying higher H₂O contents) after exposure to fluids having higher μH₂O. The behavior of natural salt-water fluid inclusions during re-equilibration was generally consistent with corresponding observations on synthetic samples verifying that compositional changes are not restricted to synthetic inclusions. Our results clearly show that there was chemical communication between fluids trapped as inclusions in quartz and the external fluid reservoir. Additionally, it is evident that although applied stress can significantly enhance the re-equilibration rate, strain in the crystal host around the inclusions resulting from large pressure differentials between the internal and confining fluids is not a necessary prerequisite for compositional change. Finally, because significant compositional changes can be induced in brine inclusions in quartz during shortterm exposure to non-equilibrium conditions at 600≤T≤825°C in the laboratory, it is likely that similar changes may result at much lower temperatures during exposure of natural rocks to non-equilibrium conditions over geologic time.     

Synthetic fluid inclusions - VII. Re-equilibration of fluid inclusions in quartz during laboratory-simulated metamorphic burial and uplift

ABSTRACT P-T conditions inferred from fluid inclusions in metamorphic rocks often disagree with the values predicted from mineral equilibria calculations. These observations suggest that inclusions formed during early stages of regional metamorphism continue to re-equilibrate during burial and subsequent uplift in response to differential pressure. P-T conditions accompanying burial and uplift were experimentally simulated by initially forming pure H₂O inclusions in quartz at elevated temperatures and pressures, and then re-equilibrating the inclusions in the presence of a 20 wt% NaCl solution such that final confining pressures ranged from 5 kbar above to 4 kbar below the initial internal pressure of the inclusions at the temperature of re-equilibration. In all samples re-equilibrated at confining pressures below the internal pressure, some inclusions were formed that had compositions of 20 wt% NaCl and densities in accord with the final P-T conditions. Additionally, some inclusions were observed to contain fluids of intermediate salinities (between 0 and 20 wt% NaCl). Densities of these inclusions were also consistent with formation at the re-equilibration P-T conditions. The remainder of the fluid inclusions observed in these samples contained pure H₂O and their homogenization temperatures corresponded to densities intermediate between the initial and final P-T conditions. In short-term experiments (7 days) where the initial internal overpressure exceeded 1 kbar, no inclusions were found that contained the original density and none were found to have totally re-equilibrated. Instead, most H₂O inclusions re-equilibrated until their internal pressures were between ∼750 and 1500 bars above the confining pressure, regardless of the initial pressure differential. In a long-term experiment (52 days), inclusions re-equilibrated at a lower confining pressure than the initial internal pressure displayed homogenization temperatures corresponding to a range in final internal pressures between 0 kbar (i.e. total re-equilibration) and 1.2 kbar above the confining pressure. In experiments where the confining pressure during re-equilibration exceeded the initial internal pressure, densities of pure H₂O inclusions increased to values intermediate between the initial and final P-T conditions. Additionally, these inclusions were generally surrounded by a three-dimensional halo of smaller inclusions, also of intermediate density, resulting in a texture similar to that previously ascribed to decrepitation from internal overpressure. In extreme cases where confining pressures were 4–5 kbar above the initial pressure, the parent inclusion almost completely closed leaving only the three-dimensional array of small (5 μm) inclusions, the outline of which may be several times the volume of the original inclusion. Groups of such inclusions closely resemble textures commonly observed in medium- to high-grade metamorphic rocks. Inclusions containing 10 and 42 wt% NaCl solutions trapped at 600 °c and 3 kbar were re-equilibrated at 600 °c and 1 kbar for 5 days in dry argon to evaluate the importance of H₂O diffusion as a mechanism of lowering the inclusion bulk density. Salinities of re-equilibrated inclusions obtained from freezing point depressions and halite dissolution temperatures indicate that original compositions were preserved. Density changes similar to those previously described were noted in these experiments, in inclusions showing no visible microfractures. Therefore, density variations observed in inclusions in this study, re-equilibrated under rapid deformation conditions, are considered to result from a change in the inclusion volume, without significant loss of contents by diffusion or leakage.     

Synthetic fluid inclusions. XV. TEM investigation of plastic flow associated with reequilibration of fluid inclusions in natural quartz

The nature and abundance of dislocations in quartz surrounding fluid inclusions were studied to obtain a better understanding of processes associated with fluid inclusion reequilibration. Synthetic fluid inclusions containing 10 wt% NaCl aqueous solution were formed in three samples at 700 °C and 5 kbar. One of the samples was quenched along an isochore to serve as a reference sample. The other two samples were quenched along a P-T path that generated internal pressures in excess of the confining pressure. The two samples were held at the final reequilibration P-T conditions of 625 °C and 2 kbar for 30 and 180 days, respectively. Following the experiments, microstructures associated with fluid inclusions were examined with the TEM. Quartz in healed fractures in the reference sample that was quenched isochorically shows a moderate dislocation activity. Quartz adjacent to reequilibrated fluid inclusions in the other two samples, however, showed a marked increase in dislocation activity compared to the un-reequilibrated sample. Deformation of the inclusion walls occurred anisotropically by expansion of mobile dislocations in their slip systems. Dislocation expansion was controlled by glide in the rhombohedral planes {1 0 1 1} that was restricted to narrow zones (≤3 μm) in the immediate vicinity of the fluid inclusion walls outside of the healed fracture plane. These plastic zones were observed after both short term (30 days) and long term (180 days) experiments and are attributed to hydrolytic weakening of quartz around fluid inclusions owing to diffusion of water into the quartz matrix during the experiment. The close spatial association of submicroscopic water bubbles with dislocations, and the rarity of water bubbles in the reference sample, show clearly that in both the 30 and 180 day experiments reequilibration involves water loss from the fluid inclusions. Our results indicate that synthetic fluid inclusions in this study recover (chemically and volumetrically), even at relatively fast experimental loading rates, such that internal stresses never reach the point of brittle failure. The driving force for fluid inclusion deformation involves two related mechanisms: plastic deformation of hydrolytically weakened wet quartz in the healed fracture, and water leakage associated with preexisting and strain-induced dislocations. Received: 5 May 1998 / Accepted: 10 February 2000     

Primary fluid inclusions in galena crystals. II. Chemical composition of the liquid and gas phase

The chemical composition of primary individual inclusions in large galena crystals from the Madan ore district, Bulgaria (see Pt. I), is studied in this paper. The liquid phase was analysed for Na, K, Ca, Fe, Mn, Mg, NH _inf4 ^sup+ , Pb, Cu, Zn, Al, Cl^–, F^–, SO ₄ ^2– , HCO ₃ ^– , and HS^– (+S^2–): pH was also determined. The following analytical methods were used: atomic-absorption spectrophotometry, emission spectral analysis, spectrophotometry, nephelometry, microcrystalloscopic and drop reactions. The gas phase was mass-spectrometrically analysed after opening the vacuoles in vacuum. It was found to consist of water vapour and CO₂ only. It was also found that the solutions are chloride-sodium-potassium with a total salinity of 4–4.5%. They do not contain measurable amounts of sulphur components. The CO₂/H₂O ratios vary strongly between the various vacuoles, which is explained by the crystallization in boiling solutions. The problems of the hermeticity of inclusions, the partial pressure of CO₂, the temperature of crystallization, and of the possible role of boiling in the ore-forming process are discussed.     

Quantitative trace element analysis of single fluid inclusions by proton-induced X-ray emission (PIXE): application to fluid inclusions in hydrothermal quartz

Single fluid inclusion analogues with known elemental composition and regular shape were analyzed for trace element contents by particle-induced X-ray emission (PIXE)—a nondestructive method for the analysis of single fluid inclusions—to evaluate the accuracy and detection limits of this type of analysis. Elements with concentrations of 10 to 1000 ppm were measured with average estimated relative error of ±7%. For natural fluid inclusions with 30 μm radius and 20 μm depth in quartz, the total analytical errors were estimated to be ±40% relative for Ca, ±16% for Fe, ±13% for Zn, ±12% for Sr, and ±11% for Br and Rb, by considering uncertainties in microscopic measurements of inclusion depths. Detection limits of 4 to 46 ppm for elements of mass numbers 25-50 were achieved for analyses of a spherical fluid inclusion with 30 μm radius and 20 μm depth in quartz, at an integrated charge of 1.0 μC. The trace element compositions of single fluid inclusions in a hydrothermal quartz crystal were also determined. The elemental concentrations in the inclusions varied widely: 0.2-9 wt.% for Ca and Fe, 300-8000 ppm for Mn and Zn, 40-3000 ppm for Cu, 100-4000 ppm for Br, Rb, Sr, and Pb, and less than 100 ppm for Ge. Elemental concentrations of secondary fluid inclusions on the same trail varied over an order of magnitude, even though all these inclusions were formed from the same fluid. Elemental concentrations in inclusions on the same trail are positively correlated with each other, except for Cu and Rb. Ratios of almost all elements in the inclusions on the trail were essentially unchanged; thus, the elemental ratios can provide original information on trace element compositions of a hydrothermal fluid.     

Experimental diffusion of hydrogen into synthetic fluid inclusions in quartz

Abstract Quartz-hosted, synthetic CO₂-H₂O fluid inclusions behave as open systems with respect to diffusional transfer of hydrogen during laboratory-simulated metamorphic re-equilibration at 650, 750 and 825°C and 1.5 kbar total pressure with f_O2 defined by the C-CH₄ buffer. Microthermometry and Raman spectroscopy show that the initial CO₂-H₂O inclusions become CO₂-CH₄-H₂-H_2O
inclusions after diffusive influx of hydrogen from the reducing confining medium. Measurable changes are observed in inclusion compositions after only 15 days of re-equilibration, implying significant hydrogen mobility at still lower temperatures over geological time spans. Results of synthetic inclusion re-equilibrium experiments have profound implications for the interpretation of natural fluid-inclusion data; failure to account for potential hydrogen migration in inclusions from high-temperature geological environments may lead to erroneous estimates of P-T, and/or the compositions of metamorphic fluids.     

Synthetic fluid inclusions - VI. Quantitative evaluation of the decrepitation behaviour of fluid inclusions in quartz at one atmosphere confining pressure

ABSTRACT The decrepitation behaviour of fluid inclusions in quartz at one atmosphere confining pressure has been evaluated using pure H₂O synthetic inclusions formed by healing fractures in natural quartz. Three different modes of non-elastic deformation, referred to as stretching, leakage or partial decrepitation, and total decrepitation have been observed. The internal pressure required to initiate non-elastic deformation is inversely related to inclusion size according to the equation: internal pressure (kbar) = 4.26 D^-0.423 where D is the inclusion diameter in microns. Regularly shaped inclusions require a higher internal pressure to initiate non-elastic deformation than do irregularly shaped inclusions of similar size. Heating inclusions through the α/β quartz inversion results in mechanical instability in the quartz crystal and leads to mass decrepitation of inclusions owing to structural mismatches generated by pressure gradients in the quartz around each inclusion. Long-term heating experiments (∼2 years) suggest that the internal pressure required to initiate non-elastic deformation does not decrease significantly with time and indicates that short-lived thermal fluctuations in natural systems should not alter the inclusion density and homogenization temperature. Inclusions that do exhibit decreased density (higher homogenization temperature) are, however, always accompanied by a change in shape from irregular to that of a negative crystal. Observations of this study are consistent with elasticity theory related to fracture generation and propagation around inclusions in minerals. These results indicate that an inclusion will not be influenced by a neighbouring inclusion, or other defect in the host phase, as long as the distance between the two is >2–4 diameters of the larger of the two inclusions.     

Statistical microthermometry of synthetic fluid inclusions in quartz during decompression reequilibration

Fluid inclusion microthermometric data are often reported as homogenization temperature frequency histograms. Interpretation of such histograms for a single fluid inclusion assemblage (FIA) of non-reequilibrated fluid inclusions is usually straightforward and provides an accurate determination of the original density (Th) of that FIA. However, interpretation of such histograms for reequilibrated inclusions is more problematic. Decompression experiments using synthetic inclusions in natural quartz and conducted at 2–5 kbar and 600–700 °C with a maximum internal overpressure of 2 kbar indicate that histogram shape reflects the sample's P-T history. Our results further indicate that the mean, mode, range, standard deviation, extreme values, etc., all have a significance with respect to the P-T history of the sample. Thus, a mound-shaped, unimodal histogram with low range is indicative of a nearly isochoric cooling P-T path. A unimodal histogram that is slightly skewed to the right, and with a low standard deviation but high range, results from inclusion deformation in the plastic regime (high temperature/low strain rates). Fluid inclusions deformed plastically show no correlation between size and density. Histogram outliers should not be ignored and may be used to determine an isochore that passes close to the conditions of entrapment (minimum Th) or close to the final reequilibration conditions (maximum Th). The histogram mean Th value corresponds to an isochore that represents the internal overpressure (about 1 kbar) that can be maintained over geologic time by a majority of reequilibrated fluid inclusions. A multimodal histogram with high range and high standard deviation indicates inclusion brittle deformation (low temperature/high strain environments). Fluid inclusions deformed in a brittle manner show strong positive correlation between size and density. Histograms produced in the laboratory show many similarities to histograms for natural samples, offering the hope that laboratory results may be used to interpret P-T histories of natural samples. Received: 20 May 1997 / Accepted: 3 April 1998     

Some effects of tectonic recrystallisation on fluid inclusions in vein quartz

Filling temperature data obtained from tectonic vein quartz varies according to the state of intracrystalline deformation. Strain free domains within grains exhibit abundant primary fluid inclusions, from which internally consistent temperatures are recorded. The onset of optical features associated with intracrystalline deformation by dislocation creep is accompanied by a decrease in the number of fluid inclusions and an increase of filling temperatures. At higher states of strain tectonic recrystallisation, evidenced by the formation of new subgrains, leads to the complete destruction of inclusion arrays. Empty cavities are swept out of the new grains during recrystallisation, into subgrain walls. Heterogeneous deformation of vein quartz at the intracrystalline level may be due in part to selective hydrolytic weakening in areas where fluid inclusions have leaked and thereby increased the structural water content.     

中国大陆科学钻探（CCSD）高压超高压变质岩中石英脉流体包裹体初步研究

初步的岩相学和冷热台显微测温表明，CCSDHP—UHP岩石中石英脉和后期碳酸盐脉中含有3种流体包裹体：盐水溶液(NaCl-H2O)包裹体(Ⅰ类)、NaCl-CaCl2-H20(Ⅱ类)和N2-CH4纯气相(Ⅲ类)包裹体，其中Ⅰ类盐水溶液包裹体可进一步分为中高盐度盐水溶液包裹体(Ia)、中等盐度盐水溶液包裹体(Ib)和低盐度盐水溶液包裹体(Ic)，而Ⅲ类为CCSD中首次发现：Ia、Ib和Ⅱ型流体包裹体主要以原生或假次生形式赋存在榴辉岩的石英脉或石英颗粒中，在角闪岩相的片麻岩及其石英脉中均未观察到，显示它们可能主要被捕获于榴辉岩的减压重结晶或退变质阶段，而Ic型包裹体分布广泛，表明其主体可能是在超高压变质岩折返过程的最晚阶段捕获的；N2-CH4纯气相包裹体均为原生包裹体，主要呈孤立和小群状与Ia和Ib类包裹体分布于榴辉岩中条带状石英脉中，可能主要是在高压-超高压榴辉岩相变质条件下被捕获的；CCSD榴辉岩中石英脉的主体形成于板块折返有关的减压重结晶和退变质，而片麻岩中石英脉则主要来源于角闪岩相及其后期退变质作用；榴辉岩和片麻岩中石英脉流体包裹体的组成和地球化学特征存在明显区别，但它们各自相似于其寄主岩石中的石英颗粒中包裹体，说明CCSD中HP—UHP岩石在板块折返过程中释放出的变质流体没有经过大规模的迁移。     

REE in fluid inclusions of quartz from gold deposits of north-eastern Russia

Inductively coupled plasma-mass spectrometry (ICP-MS) has been used to determine rare earth element concentrations in aqueous solutions extracted from fluid inclusions. Quartz has been sampled from ores of three major types of polygenic gold hydrothermal systems of North-Eastern Russia: (1) gold-quartz-sulphide (Au-Q, Nezhdaninsk); (2) gold-antimony (Au-Sb, Sarylakh) and (3) intrusion-related gold-bismuth-siderite-polysulphide (Au-Bi-Sid, Arkachan) large deposits located in terrigenous rocks of the Verkhoyansk fold belt. The total concentration of REE in the fluid inclusions is not high (up to 52 ppm). The contribution of LREE dominates in REE balance (??LREE/??HREE=7.4?C112.1). The chondrite-normalized REE patterns of inclusion fluids for the Au-Q and Au-Bi-Sid deposits are characterized by LREE enrichment with a positive or negative Eu anomaly. REE patterns for the regenerated quartz from Au-Sb deposits are characterized by pronounced differentiation between light and heavy lanthanides in fluid inclusions. Significant total REE concentration decreasing (on 1?C2 order) from early to late stages of Nezhdaninsk and Arkachan deposits is revealed. The positive correlations of total REE concentrations with Rb, Cs, Li and B contents in fluid inclusions are shown. The REE distribution in fluid inclusions can be used as indicators of the contribution of magmatic fluid in the hydrothermal system.     

Salinity of the Archaean oceans from analysis of fluid inclusions in quartz

Fluids trapped in inclusions in well-characterized Archaean hydrothermal quartz crystals were analyzed by the extended argon–argon method, which permits the simultaneous measurement of chlorine and potassium concentrations. Argon and nitrogen isotopic compositions of the trapped fluids were also determined by static mass spectrometry. Fluids were extracted by stepwise crushing of quartz samples from North Pole (NW Australia) and Barberton (South Africa) 3.5–3.0-Ga-old greenstone belts. The data indicate that fluids are a mixture of a low salinity end-member, regarded as the Archaean oceanic water, and several hydrothermal end-members rich in Cl, K, N, and radiogenic parentless ⁴⁰Ar. The low Cl–K end-member suggests that the salinity of the Archaean oceans was comparable to the modern one, and that the potassium content of the Archaean oceans was lower than at present by about 40%. A constant salinity of the oceans through time has important implications for the stabilization of the continental crust and for the habitability of the ancient Earth.     

Part II. Evaluation of Ar-Ar quartz ages: Implications for fluid inclusion retentivity and determination of initial Ar/Ar values in Proterozoic samples

The argon isotope systematics of vein-quartz samples with two different K-reservoirs have been evaluated in detail. Potassium is hosted by ultra-high-salinity fluid inclusions in quartz samples from the Eloise and Osborne iron-oxide-copper-gold (IOCG) deposits of the Mt Isa Inlier, Australia. In contrast, K is hosted by accidentally trapped mica within lower-salinity fluid inclusions of a sample selected from the Railway Fault, 13 km south of the Mt Isa copper mine, Australia. Imprecise apparent ages have been obtained for all of the samples studied and conclusively demonstrate that quartz fluid inclusions are retentive to Ar and have not leaked over billions of years. IOCG samples that host K in fluid inclusions only, have K/Cl values of <1 and the ages obtained represent the maximum ages for mineralization. In contrast, the Railway Fault samples that include accidentally trapped mica have K/Cl values of ?1. Excess ⁴⁰Ar_E plus Cl hosted by fluid inclusions, and radiogenic ⁴⁰Ar_R plus K, are strongly correlated in these samples and define a plane in 3D ⁴⁰Ar-³⁶Ar-K-Cl space. In this case, the plane yields an ‘excess ⁴⁰Ar_E’ corrected age of ∼1030 Ma that is 100’s of Ma younger than nearby Cu-mineralization at Mt Isa. The age is interpreted to reflect ⁴⁰Ar-loss from the accidentally trapped mica into the surrounding fluid inclusions, and is not related to the samples’ age of formation. The initial ⁴⁰Ar/³⁶Ar value of fluid inclusions is widely used to provide information on fluid origin. For the IOCG samples that host K in fluid inclusions only, the initial ⁴⁰Ar/³⁶Ar values are close to the measured values at every temperature of stepped heating experiments. For samples that include accidentally trapped mica, the correction for post-entrapment radiogenic ⁴⁰Ar_R production is significant. Furthermore, because ³⁹Ar_K present in accidentally trapped mica crystals is released at different temperatures to radiogenic ⁴⁰Ar_R lost to the surrounding fluid inclusions, intra-sample ⁴⁰Ar/³⁶Ar variation cannot be reliably documented. The results demonstrate that noble gas analysis is readily applicable to Proterozoic, or older, samples but that if K-mineral impurities are present within quartz the abundance of K must be determined before calculation of mean ⁴⁰Ar/³⁶Ar values that are representative of the samples’ initial composition.