The electrode-electrolyte interface was also investigated in this study. We used the technique of impedance spectroscopy to analyse this interface. Both charge transfer and diffusion coefficient have been determined from the Nyquist diagrams. The values obtained were in good agreement with those obtained from the above proposed model.

This structure is compared to those of the A(II)M(III)F5 compounds (BaGaF5, BaAlF5, BaMnF5, SrNiF5).

Further insight is gained into the nature of the processes occurring during the cure of the matrix by a comparison of dielectric data with mechanical and thermal data obtained using a curometer. A good correlation appears between the various transition points identified from the different measurements. Comparison of the data from the various methods and also that from previous workers is made using a TTT diagram.

The growth kinetics of 35 pleuridia were analysed using Nelder's generalized logistics. Each sigmoidal curve, which was divided into four growth stages from the instantaneous acceleration variations, was thus characterized by ten kinetic parameters: lengths at the time of the changes in growth stage, durations of the growth stages and maximum growth rate.

The data showed that all curves displayed a marked left dissymmetry which reflected the brevity of the accelerated growth stage (1/5 of total growth duration). The maximum acceleration stage occurs between the individualization of the pleuridial initial cell and the first mitosis. Previously known experimental data were related to this quantitative result so as to put forward a biological meaning. New research pathways were suggested.

The kinetic parameters investigated were used for a classification of pleuridia. The smallest ones displayed the slowest growth: they needed twice more time to reach their maximum size than the largest pleuridia whose growth was the fastest. This association of traits was related to the occurrence of acropetal fluxes along the main axis of the thallus.

The use of mathematical theories makes it possible to state the hypotheses under which reverberation is defined. These hypotheses amount to two: the limit of high frequencies, necessary to define sound rays; and diffusion. The latter is redefined as equivalent to mixing, together with a method for computing a diffusion index: the mixing time. Finally, the consequences of relaxing the hypothesis of diffusion are analysed.

A series of tests was set up to check whether the "35 ms fusion" rule still holds in the presence of background noise. The listening conditions are similar to those found in a car cabin (low-frequency noise, strong reflection arriving within 10 ms). The tests reveal that the actual word spoken and the speaker's voice dominate over other sources of influence on the results, even for a reflection following 5 ms after the direct sound. Although its energy is actually integrated with that of the direct sound, an early reflection never improves the intelligibility to the same extent as simply raising the voice to the same level.

A second series of tests was therefore conducted to investigate the word effect: it was found that this effect is also listener dependent, but no physical correlate could explain it. Nevertheless, the experiments agree in showing that the "fusion" rule does not hold in the presence of background noise.

Methods. So we decided in 1984 to test a new multiple drug procotol (Tours-Poitiers-Limoges = TPL protocol) which used a sequence of three courses of classical high-dose induction therapy, three courses of consolidation therapy using Teniposide, Cytosine Arabinoside, L Asparaginase and high-dose Methotrexate, and three courses of late intensification using the same drugs as induction therapy.

Results. Thirty-eight patients younger than 60 years were included. Complete remission was obtained in 27 patients (71%). The median follow-up was 3 years and 9 months with one third of CR patients having been followed beyond 5 years. Seven patients relapsed (26% of CR patients) and one died of toxicity in complete remission. At present 22 patients (58%) are in complete remission, 19 in first CR, 1 in first CR after allogenic bone marrow transplantation, and 2 in second prolonged CR after autologous bone marrow transplantation. The median survival time is 48 months and the actuarial disease-free survival curve seems to have a plateau at 48.5%, with no relapse after 24 months.

Conclusions. These results confirm the efficacy of alternating high-dose conventional chemotherapy in the treatment of intermediate and high-grade NHL, with about half of the patients being cured. However, more intensive chemotherapy regimens are needed to improve cure rates.

Results of the subjective tests carried out in the theatre are looked at in terms of their correlations with objective measurements, and observations on the relevance of such tests for characterising a hall are derived. Specially, we believe the audition-vision interaction can't be ignored.

The powder diffraction patterns are close to that of Li2TiTeO6. For x < 0.8, the orthorhombic symmetry of the starting material and the initial Ti/Te ordering are maintained (S.G. Pnn2, Z = 2). But, for x > 0.8, we observe a migration of Ti atoms in a preserved anionic network, involved by the departure of the last Li atoms.

In a first step, the structure get settled in a trirutile system with cationic vacancies (centric space group Pnnm). Then, a concerted migration of Ti atoms, obtained by further heating, gives an ordered trirutile phase (acentric space group P42nm).

The structures are refined by a modified Rietveld method applied to the X-Ray diffraction patterns. This study illustrates a new example of topotactic mechanism involved in the Li+/H+ exchange reactions on complex oxides.