_Melting Point._--This is best determined by melting some of the solid oil, or crystals, and sucking a small quant.i.ty up into a capillary tube, which is then attached by a rubber band to the bulb of the thermometer, immersed in a suitable bath (water, glycerine, oil, etc.) and the temperature of the bath gradually raised until the substance in the tube is sufficiently melted to rise to the surface, the temperature at which this takes place being the melting point.

The melting point of otto of rose is usually taken in a similar tube to the setting point, and is considered to be the point at which the last crystal disappears.

_Iodine Absorption._--In the authors' opinion, this is of some value in conjunction with other data in judging of the purity of otto of rose. It is determined by Hubl's process as described under Fats and Oils, except that only 0.1 to 0.2 gramme is taken, and instead of 10 c.c. of chloroform, 10 c.c. of pure alcohol are added. The rest of the process is identical.

SOAP.

In the a.n.a.lysis of soap, it is a matter of considerable importance that all the determinations should be made on a uniform and average sample of the soap, otherwise very misleading and unreliable figures are obtained.

Soap very rapidly loses its moisture on the surface, while the interior of the bar or cake may be comparatively moist, and the best way is to carefully remove the outer edges and take the portions for a.n.a.lysis from the centre. In the case of a household or unmilled toilet soap, it is imperative that the quant.i.ties for a.n.a.lysis should all be weighed out as quickly after each other as possible.

_Fatty Acids._--Five grammes of the soap are rapidly weighed into a small beaker, distilled water added, and the beaker heated on the water bath until the soap is dissolved.

A slight excess of mineral acid is now added, and the whole heated until the separated fatty acids are perfectly clear, when they are collected on a tared filter paper, well washed with hot water and dried until constant in weight. The result multiplied by 20 gives the percentage of fatty acids in the sample.

A quicker method, and one which gives accurate results when care is bestowed upon it, is to proceed in the manner described above as far as the decomposition with mineral acid, and to then add 5 or 10 grammes of stearic acid or beeswax to the contents of the beaker and heat until a clear layer of fatty matter collects upon the acid liquor.

Cool the beaker, and when the cake is sufficiently hard, remove it carefully by means of a spatula and dry on a filtering paper, add the portions adhering to the sides of the beaker to the cake, and weigh.

The weight, less the amount of stearic acid or beeswax added, multiplied by 20 gives the percentage of fatty acids.

Care must be taken that the cake does not contain enclosed water.

The results of these methods are returned as fatty acids, but are in reality insoluble fatty acids, the soluble fatty acids being generally disregarded. However in soaps made from cocoa-nut and palm-kernel oils (which contain an appreciable quant.i.ty of soluble fatty acids) the acid liquor is shaken with ether, and, after evaporation of the ethereal extract, the amount of fatty matter left is added to the result already obtained as above, or the ether method described below may be advantageously employed.

Where the soap under examination contains mineral matter, the separated fatty acids may be dissolved in ether. This is best performed in an elongated, graduated, stoppered tube, the total volume of the ether, after subsidence, carefully read, and an aliquot part taken and evaporated to dryness in a tared flask, which is placed in the oven at 100 C. until the weight is constant.

In a complete a.n.a.lysis, the figure for fatty acids should be converted into terms of fatty anhydrides by multiplying by the factor 0.9875.

In this test the resin acids contained in the soap are returned as fatty acids, but the former can be estimated, as described later, and deducted from the total.

_Total Alkali._--The best method is to incinerate 5 grammes of the soap in a platinum dish, dissolve the residue in water, boil and filter, making the volume of filtrate up to 250 c.c., the solution being reserved for the subsequent determination of salt, silicates, and sulphates, as detailed below.

Fifty c.c. of the solution are t.i.trated with N/1 acid, to methyl orange, and the result expressed in terms of Na_{2}O.

Number of c.c. required 0.031 100 = per cent. Na_{2}O.

The total alkali may also be estimated in the filtrate from the determination of fatty acids, if the acid used for decomposing the soap solution has been measured and its strength known, by t.i.trating back the excess of acid with normal soda solution, when the difference will equal the amount of total alkali in the quant.i.ty taken.

The total alkali is usually expressed in the case of hard soaps as Na_{2}O, and in soft soaps as K_{2}O.

_Free caustic alkali_ is estimated by dissolving 2 grammes of the soap, in neutral pure alcohol, with gentle heat, filtering, well washing the filter with hot neutral spirit, and t.i.trating the filtrate with N/10 acid, to phenol-phthalein.

Number of c.c. required 0.0031 50 = per cent. free alkali Na_{2}O, as caustic.

_Free Carbonated Alkali._--The residue on the filter paper from the above determination is washed with hot water, and the aqueous filtrate t.i.trated with N/10 acid, using methyl orange as indicator. The result is generally expressed in terms of Na_{2}O.

Number of c.c. required 0.0031 50 = per cent. free alkali Na_{2}O, as carbonate.

_Free Alkali._--Some a.n.a.lysts determine the alkalinity to phenol-phthalein of the alcoholic soap solution without filtering, and express it as free alkali (caustic, carbonates, or any salt having an alkaline reaction).

_Combined Alkali._--The difference between total alkali and free alkali (caustic and carbonate together) represents the alkali combined with fatty acids. This figure may also be directly determined by t.i.trating, with N/2 acid, the alcoholic solution of soap after the free caustic estimation, using lacmoid as indicator.

The potash and soda in soaps may be separated by the method described for the estimation of pota.s.sium in _Pearl ash_ (page 126).

The pota.s.sium platino-chloride (K_{2}PtCl_{6}) is calculated to pota.s.sium chloride (KCl) by using the factor 0.3052, and this figure deducted from the amount of mixed chlorides found, gives the amount of sodium chloride (NaCl), from which the sodium oxide (Na_{2}O) is obtained by multiplying by 0.52991.

The pota.s.sium chloride (KCl) is converted into terms of pota.s.sium oxide (K_{2}O) by the use of the factor 0.63087.

_Salt_ may be determined in 50 c.c. of the filtered aqueous extract of the incinerated soap, by exactly neutralising with normal acid and t.i.trating with N/10 silver nitrate solution, using a neutral solution of pota.s.sium chromate as indicator. The final reaction is more distinctly observed if a little bicarbonate of soda is added to the solution.

Number of c.c. required 0.00585 100 = per cent. of common salt, NaCl.

Chlorides may also be estimated by Volhard's method, the aqueous extract being rendered slightly acid with nitric acid, a measured volume of N/10 silver nitrate solution added, and the excess t.i.trated back with N/10 ammonium thiocyanate solution, using iron alum as indicator.

_Silicates._--These are estimated by evaporating 50 c.c. of the filtered extract from the incinerated soap, in a platinum dish with hydrochloric acid twice to complete dryness, heating to 150 C., adding hot water, and filtering through a tared filter paper.

The residue is well washed, ignited, and weighed as SiO_{2}, and from this silica is calculated the sodium silicate.

_Sulphates_ may be determined in the filtrate from the silica estimation by precipitation with barium chloride solution, and weighing the barium sulphate, after filtering, and burning, expressing the result in terms of Na_{2}SO_{4} by the use of the factor 0.6094.

_Moisture._--This is simply estimated by taking a weighed portion in small shavings in a tared dish, and drying in the oven at 105 C. until it ceases to lose weight. From the loss thus found is calculated the moisture percentage.

_Free or Uncombined Fat._--This is usually determined by repeated extraction of an aqueous solution of the soap with petroleum ether; the ethereal solution, after washing with water to remove traces of soap, is evaporated to dryness and the residue weighed.

A good method, which can be recommended for employment where many determinations have to be performed, is to dissolve 10 grammes of soap in 50 c.c. neutral alcohol and t.i.trate to phenol-phthalein with N/1 acid. Add 3-5 drops HCl and boil to expel carbonic acid, neutralise with alcoholic KOH solution and add exactly 10 c.c. in excess, boil for fifteen minutes under a reflux condenser and t.i.trate with N/1 acid. The difference between this latter figure and the amount required for a blank test with 10 c.c. alcoholic KOH, denotes the amount of alkali absorbed by the uncombined fat.

_Examination of the fatty acids_ as a guide to the probable composition of the soap:--

From the data obtained by estimating the "t.i.tre," iodine number, and saponification equivalent of the mixed fatty and rosin acids, and the rosin content, a fairly good idea of the const.i.tution of the soap may be deduced.

The t.i.tre, iodine number, and saponification equivalent are determined in exactly the same manner as described under Fats and Oils.

The presence of rosin may be detected by the Liebermann-Storch reaction, which consists in dissolving a small quant.i.ty of the fatty acids in acetic anhydride, and adding to a few drops of this solution 1 drop of 50 per cent. sulphuric acid. A violet coloration is produced with rosin acids. The amount of rosin may be estimated by the method devised by Twitch.e.l.l (_Journ. Soc. Chem. Ind._, 1891, 804) which is carried out thus:--

Two grammes of the mixed fatty and rosin acids are dissolved in 20 c.c.

absolute alcohol, and dry hydrochloric acid gas pa.s.sed through until no more is absorbed, the flask being kept cool by means of cold water to prevent the rosin acids being acted upon. The flask, after disconnecting, is allowed to stand one hour to ensure complete combination, when its contents are transferred to a Philips' beaker, well washed out with water so that the volume is increased about five times, and boiled until the acid solution is clear, a fragment of granulated zinc being added to prevent b.u.mping. The heat is removed, and the liquid allowed to cool, when it is poured into a separator, and the beaker thoroughly rinsed out with ether. After shaking, the acid liquor is withdrawn, and the ethereal layer washed with water until free from acid. Fifty c.c. neutral alcohol are added, and the solution t.i.trated with N/1 KOH or NaOH solution, the percentage of rosin being calculated from its combining weight. Twitch.e.l.l suggests 346 as the combining weight of rosin, but 330 is a closer approximation.

The method may be also carried out gravimetrically, in which case petroleum ether, boiling at 74 C. is used for washing out the beaker into the separator. The acid liquor is run off, and the petroleum ether layer washed first with water and then with a solution of 1/2 gramme KOH and 5 c.c. alcohol in 50 c.c. water, and agitated. The rosin is thus saponified and separated. The resinate solution is withdrawn, acidified, and the resin acids collected, dried and weighed.

_Halphen's Reaction._--This is a special test to determine the presence or absence of cotton-seed oil fatty acids in mixtures. Equal parts of the fatty acids, amyl alcohol, and a 1 per cent. solution of sulphur in carbon bisulphide, are heated in a test-tube placed in a water-bath until effervescence ceases, then in boiling brine for one hour or longer when only small quant.i.ties are present. The presence of cotton-seed oil is denoted by a pink coloration. The reaction is rendered much more rapid, according to Rupp (_Z. Untersuch. Nahr. Genussm._, 1907, 13, 74), by heating in a stoppered flask.

Other bodies which it is occasionally necessary to test for or determine in soap include:--

_Carbolic acid._--Fifty grammes of the soap are dissolved in water and 20 c.c. of 10 per cent. caustic potash added. The solution is treated with an excess of brine, the supernatant liquor separated, and the precipitate washed with brine, the washings being added to the liquor withdrawn. This is then evaporated to a small bulk, placed in a Muter's graduated tube, and acidified with mineral acid.

The volume of separated phenols is observed and stated in percentage on the soap taken.

Or the alkaline layer may be rendered acid and steam distilled; the distillate is made up to a known volume, and a portion t.i.trated by the Koppeschaar method with standard bromine water.