On the Fixation of Atmospheric Nitrogen by Phoma Radicis Callunae, Including A New Method for Investigating Nitrogen-Fixation in Micro-Organisms

It is a well-established fact that the bacteria occurring in the root-nodules of Leguminous plants have the power of “fixing” atmospheric nitrogen, and the suggestion that a similar capacity is possessed by the fungi regularly associated in the form of mycorrhiza with the roots of higher plants is not a novel one.

In the two families Orchidaceae and Ericaceae there is an obligate relation between the higher plant and the endophytic fungus concerned. It is natural, therefore, that in the investigation of the nitrogen-fixing capacity of mycorrhizal fungi, special attention should be paid to the members of these two families and to their specific fungi.

Earlier experimental researches on the root fungi of Orchids yielded negative results for nitrogen-fixation, but the recent work of Wolff (1926) points to a considerable degree of fixation in the root fungus of Neottia and other genera. A list of the literature dealing with nitrogen-fixation in the Orchidaceae will be found at the end of his paper¹. The present research is concerned with nitrogen-fixation in Calluna vulgaris, a member of the Ericaceae.

Ternetz (1904, 1907) was the first to isolate the mycorrhizal fungi of Ericaceae and to grow them in pure culture. From the roots of five ericaceous species, she isolated fungi which she believed to be those responsible for the formation of mycorrhiza in these species. Subsequent work by Rayner (1915) has substantiated her belief. Ternetz investigated the behaviour of these fungi when grown on various media, with a view to determining their capacity for fixing atmospheric nitrogen. As a result, she concluded that all the fungi isolated could utilise the nitrogen of the atmosphere in varying degrees.

Table I, below, derived from Table IX in Ternetz’s paper (1907), gives a summary of her results :

Ternetz employed for her work the method in general use for this kind of investigation. Air, freed from gaseous nitrogenous compounds and filtered through cotton-wool, is bubbled through the liquid culture and the nitrogen content of the latter is determined at the end of the growth period. This method is open to errors arising from contamination of the culture (1) by organisms in the atmosphere, (2) by nitrogenous compounds in the air, which are only incompletely removed by bubbling the air slowly through wash-bottles of sulphuric acid and sodium hydrate. The other main source of error is in the final estimation of the nitrogen in the cultures. The first source of error can be checked by a microscopic examination of the cultures after growth. This was done by Ternetz throughout her work, but although information is thus gained regarding this source of error, the error itself cannot be entirely avoided. As regards the second source of error, we have found on trial that the last traces of substances such as ammonia and carbon dioxide are not always removed from air by slowly bubbling it through absorptive solutions of concentrated sulphuric acid or sodium hydrate respectively. When air was slowly sucked through a 1-inch depth of concentrated sulphuric acid in a wash-bottle and then through a bottle containing Nessler’s solution, the latter became yellow in colour in two days. In cultures grown for periods of 3—5 weeks, the accumulation of ammonia and other compounds might be significant and would then vitiate any positive result as to nitrogen-fixation on the part of the fungus.

In a paper published by B. M. Duggar and A. R. Davis (1916) on the fixation of nitrogen by fungi, the authors detail their own results and give a critical review of all the available literature relating to the subject. In the course of this review, the technique employed by Ternetz is closely scrutinised. It is considered that the utmost care was taken with regard to the purity of materials, the use of necessary blanks in the analyses and of controls in the experiments. It is pointed out, however, that only from one-eighteenth to one-fourth of the solution was taken for analysis in each case, so that any small experimental error is magnified 4—18 times. This is the only serious criticism brought against the methods of Ternetz.

We would like to point out, however, that although Ternetz quotes figures for nitrogen estimations to four places of decimals in milligrams, i.e. to 10⁻⁷ gm., her technique does not justify this. For instance, in her nitrogen analyses, Ternetz uses N/10 solutions of acid and alkali for titration and quotes her burette readings to 0.05 c.c. Now 0.05 c.c. N/10 solution = 0.00007 gm nitrogen, so that on this determination alone the results are certainly not more accurate than 0.00007 gm. Generally in such determinations, several analyses of similar type are performed and the average value has then a far smaller probable error than the individual values. But it does not appear from her paper that Ternetz made use of the one very great advantage of the method of analysis of aliquot parts she employed, namely that in this method it is possible to carry out duplicate determinations and so obtain a check on the results. So that the accuracy of her figures does not exceed the accuracy of the individual Kjeldahl analyses. This by no means detracts from the value of the work so long as it is borne in mind that the last three figures have no real meaning. Another rather similar criticism of Ternetz’s work is that she bases her conclusions on the results of single cultures, never, apparently, setting up her cultures in duplicate.

Although no mycorrhizal fungus was investigated by Duggar and Davis, they included in their work Phoma betae…a member of the same genus as the root fungi of ericaceous species. It is noteworthy that Phoma betae is conspicuous among the fungi investigated for its high nitrogen-fixing capacity. In fact, Duggar and Davis ascribe definite nitrogen-fixation to this species alone.

Such methods of direct chemical estimation of nitrogen-fixation have not been applied hitherto to Phoma radicis Callunae, the mycorrhizal fungus of Calluna vulgaris; although the results obtained by Ternetz with root fungi of other Erica ceous genera, and by Duggar and Davis with the parasitic Phoma betae suggest the likelihood of nitrogen-fixation in this species of Phoma also.

Indirect evidence in favour of this view has already been supplied by Rayner (1922), who succeeded in germinating sterile seeds of Calluna vulgaris, inoculating them with a pure culture of the endophyte and carrying on the resulting seedlings for many months of healthy growth. The mineral salts in the culture medium were varied in different sets of experiments, and it was found that the optimum conditions for growth were provided when no source of combined nitrogen was available. The seedlings, grown for several months on a medium entirely lacking combined nitrogen, were healthy and dark green in colour, showing no sign whatever of nitrogen starvation.

It is suggested by the author that this behaviour would be accounted for if the fungus partner possessed a capacity for utilising atmospheric nitrogen, a property which would explain also the ability of Calluna to grow satisfactorily under natural conditions on soils deficient in nitrogen compounds. It is not considered likely that the experimental plants obtained sufficient nitrogenous food from traces of nitrogen compounds in the air. Doubtless such sources can be utilised, but it seems improbable that they would have provided more than an insufficient fraction of the nitrogen required by the cultures under consideration, which were grown in a cool greenhouse in the open garden of the University of Reading. Nor are the nitrogenous reserves of the seed regarded as likely to be significant in amount in view of the very small size of the seeds in this species.

Dr Rayner has been kind enough to supply a number of these synthetic seedlings of Calluna grown on nitrogen-free media. They have been analysed for nitrogen content for comparison with the nitrogen content of ungerminated seed.

The scope of the present research can be summarised as follows :

In view of the difficulties in obtaining conclusive evidence of nitrogen-fixation on the part of micro-organisms by the methods usually employed, an alternative method has been devised : this method is described.

In the second place, since a pure culture of the mycorrhizal fungus of Calluna vulgaris was available, it seemed desirable to obtain direct chemical data, hitherto wanting, as to the nitrogen-fixing capacity of this fungus. An attempt to provide such data has been made by means of the new method referred to above.

Finally, the evidence in favour of nitrogen-fixation by Calluna seedlings brought forward by Rayner has been put on a quantitative chemical basis.

As mentioned at the end of the introductory section, synthetic seedlings of Calluna were observed to grow vigorously on a nitrogen-free medium¹, showing no signs of nitrogen starvation. The only available sources of nitrogenous supply for such seedlings are the reserves in the seed and the surrounding atmosphere. If the seedlings contain markedly larger amounts of nitrogen than do the seeds, this must have been derived from the latter source.

The seedlings were some of a batch grown on agar in large test-tubes during 1916, since when the test-tubes had been filled with alcohol, corked up and waxed over the corks. The agar medium in which the seedlings were grown had been previously analysed and proved to contain no detectable nitrogen.

The results of analyses of the seeds and seedlings are given in the following table. (For the Method of Analysis, see p. 177.)

In considering the results shown in Table II, it is noteworthy that seedling a was a good deal larger than the other three seedlings of which d was the smallest.

The probable error of each experiment is computed at 0.00007 gm. (see p. 181) so that the values for nitrogen content of the seedlings are rather larger than the experimental error.

From the results of the analyses of the Calluna seeds it will be seen that individual seeds contain no detectable nitrogen, the value being of the order of 10⁻⁶ gm. per seed.

Nitrogen must therefore have been obtained by the seedling from some source external to the seed and since no nitrogen compounds were supplied in the medium, it is suggested that this source was the molecular nitrogen of atmosphere since, as was pointed out above, it seems unlikely that, under the conditions of growth, sufficient compounds of nitrogen would have been available from the air, not merely to prevent the appearance of symptoms of nitrogen hunger, but to promote vigorous growth.

The method about to be described is suitable for the investigation of nitrogenfixation in any micro-organisms which will grow on a liquid culture medium. The culture is grown in an apparatus entirely closed to the external atmosphere. The advantages of this method over the older ones are that, since the same air circulates during the whole period of growth and no fresh air need ever be admitted to the system, risk of contamination by nitrogen compounds from the air is very much diminished and may in fact be said to be avoided completely. Moreover, when air is drawn through cultures continuously for a period of several weeks, there is risk of contamination by micro-organisms : such a risk is obviated in the present method. Furthermore, the new method makes it possible to grow a control culture in an atmosphere containing no nitrogen in any form whatsoever, whether combined, or as the molecular nitrogen of the atmosphere. Thus, at the end of the growth period, instead of comparing the nitrogen content of two culture flasks, one of which has been inoculated and the other not, comparison is made between two flasks both of which have been inoculated, but one of which contains gaseous nitrogen and the other is without this possible source of food supply.

The apparatus is composed of a number of similar units, the parts constituting a single unit being shown in Fig. 1. The system can be broken at X and Y, and the portions between can be autoclaved in one piece and rejoined to the rest of the apparatus by rubber tubing. In practice it was found possible to operate six such units from the same pump by means of T-pieces inserted at V ; and by insertion of T-pieces at A, one voltameter was made to serve three units containing nitrogen, while another served three units containing no nitrogen—two voltameters thus sufficing for six units. The Plate shows three units assembled, that is half the apparatus as actually used.

The fungus is grown in the culture flask A, which is connected with a U-tube B. The latter contains broken pumice soaked in concentrated NaOH (made up with ammonia-free water) to absorb the CO₂ respired by the fungus, and a rolled sheet of gold leaf is also inserted in one of the limbs of the U-tube to absorb any mercury vapour which might be carried over from the pump and might be toxic to the organism. The other arm of the U-tube is connected with a mercury valve system after the design of Blackman and Bolas (1926), the closed circuit being completed as shown. D is a two-way tap which serves to admit either gases from the voltameter or sterilised air through the side arm ; F is the outlet tap which is opened only while gases are being passed into the apparatus through the tap D; E is open during the main part of the experiment, but closed while voltameter gases or sterile air are being passed through the apparatus.

The voltameter is of the form shown on the left of Fig. 1, and is run from the lighting main through a variable mat resistance. The two-way taps G and H allow the gas from either electrode to escape to the external atmosphere or to be passed through the rest of the apparatus at will. J is a glass tube containing MnO₂ to decompose any ozone which might be formed as a product of electrolysis. [In point of fact, pure baryta (provided that it does not contain much carbonate) gives a very pure electrolytic gas, free from ozone, hydrogen peroxide and hydrocarbons. In the apparatus now under consideration the only exit to the atmosphere is closed by a tube containing soda-lime, while any barium carbonate formed being insoluble sinks to the bottom and does not come into contact with the electrodes.]

The portion between G, H and X can be detached and autoclaved in one piece. The burettes of the voltameter are kept sterile by their contact with the concentrated baryta, which can be sucked right to the top before the experiment begins. Thus the whole apparatus can be rendered aseptic.

The experiment is set going in the following way: All portions of the apparatus are thoroughly cleansed with chromic acid, washed out with distilled water and finally with nitrogen-free water (redistilled from acid permanganate). Into the pump is introduced a requisite quantity of clean mercury, and into B the NaOH and gold leaf as described. Into A the required volume of the culture solution is carefully measured by means of a pipette. The taps are greased with rubber grease and a thin thread of cotton-wool put through the aperture before replacing the stopper in order that the tap may not stick during the autoclaving. The exterior of the taps and all joints and exit tubes are then bound with cotton-wool. The tube leading into the flask A is withdrawn a little through the rubber stopper until it is above the level of the liquid in the flask and the whole autoclaved at 15 lbs. pressure for 20 minutes¹. When cool, the flask is inoculated with careful aseptic precautions, the tube slid through its stopper until it dips below the surface of the culture liquid, and connections made with the rest of the apparatus which has been previously sterilised.

The unit is then supplied with its artificial atmosphere, containing or not containing nitrogen, as the case may be. This is performed as follows : The taps D and F are opened and E is closed, the current is switched on and the pump is set in motion so that the unit is swept out with electrolytic gas. After six hours, in which time the Original gases have all been removed (see Appendix), the current is switched off, D and F are closed and E opened, and the pump which had been momentarily stopped is started again. The gas in the apparatus, which contains no nitrogen in any form, now circulates in the direction of the arrows. In the cultures to which gaseous nitrogen is to be supplied, the unit is swept through with oxy-hydrogen gas for two hours only, since here it is not necessary to remove all traces of atmospheric nitrogen, and then a system for supplying sterilised air is attached at D. This consists of a spiral of copper tubing which can be heated in a bunsen flame, connected to a cooling coil immersed in cold water, which is attached in series with two wash-bottles containing H₂SO₄ and NaOH solutions respectively. (Before attachment two or three litres of air should be drawn through the system to ensure the sterility of all the air inside.) E is then closed and F opened, while D is opened to the side arm. When the pump is set going it will therefore introduce sterilised air into the unit, the volume of which can be measured by collecting the gas issuing from F. (The rate of the air stream through the apparatus, and therefore the rate of bubbling through the H₂SO₄ and NaOH solutions, is quite slow, and the total volume of air introduced is small ; hence the risk that the air is contaminated with significant amounts of nitrogen compounds is extremely small.) When 150 c.c. of sterilised air has been admitted, the three stopcocks are adjusted as before and the apparatus for supplying sterilised air is detached.

The object in thus filling all the units with electrolytic gas, whether or no they are ultimately to contain gaseous nitrogen and then admitting sterilised air to one series of units, is to obtain a perfect control, in the event of certain constituents of the voltameter being harmful to growth. In the later experiments, when it had been proved that the cultures in the oxy-hydrogen gas grew as well as those in air, and hence that the gas supplied from the voltameter could have no such ill effects, the above procedure was discontinued and the units that were ultimately to contain air were simply autoclaved, inoculated and connected to the rest of the apparatus and the pump was started up without opening the taps D and F. In these experiments even the slight risk of contamination by nitrogenous compounds or by organisms while supplying the apparatus with its artificial atmosphere was avoided. The fact that only one out of the whole series of cultures grown over a period of 18 months was contaminated after four weeks’ growth testifies to the efficacy of the method for providing and maintaining aseptic conditions. At intervals any deficiency in oxygen due to respiration of the fungus is made good by momentarily adjusting the appropriate taps so as to put the unit into communication with the oxygen side of the voltameter. At intervals also, if so desired, the atmospheres in any of the units can be completely replaced without difficulty and without fear of casual contamination¹.

The flask A 1 (No. 1) containing nitrogen gas and its control A 2 (No. 1) without nitrogen gas are kept immersed in the same water bath¹ so that they are always at the same temperature (and similarly for A 1 (No. 2) and A 2 (No. 2) and for the other pairs), but they are not connected to the same voltameter. The voltameter connections are arranged so that all the flasks containing nitrogen gas are served by one voltameter, and all the flasks containing no nitrogen gas by another, so as to avoid any possibility of leakage of nitrogen gas to the control flasks from those supplied with nitrogen.

A form of Phoma radicis was extracted from roots of Calluna vulgaris by Ternetz (1907) but was not investigated for nitrogen-fixing capacity. The identity of this form with the endophyte was confirmed by Rayner (1915).

The fungus isolated by Rayner was extracted from seeds removed with aseptic precautions from unopened fruits and its identity proved by re-inoculation into seedlings grown in pure culture. In pure culture it agreed in all respects with that originally obtained by Ternetz.

The strain used in the present research was re-isolated by Rayner from seeds in the early spring of 1924 and had therefore been cultivated outside the plant on various organic media for two or three years subsequent to extraction.

When first isolated it grew vigorously on many different media, producing pycnidia and spores freely. Six months after extraction capacity to form pycnidia disappeared and was never regained in spite of many changes to media of different constitution. The vegetative vigour declined steadily and was markedly less at the time it was used than when first isolated.

The loss of sporing capacity under laboratory culture has been observed also in other species of Phoma (Westerdijk, J. and Luijk, A. van 1920).

In the first series of cultures (Series A) a culture medium similar to that used by Ternetz and having the following composition was employed:

This solution has a hydrogen-ion concentration of pH 5.0 (approx.).

All reagents used were A.R. brand of the British Drug Houses and all solutions were prepared with glass-distilled water, redistilled from acid permanganate and tested for ammonia with Nessler’s reagent.

In the earlier experiments 50 c.c. of culture fluid was used in each experiment. Later it was thought that the observed arrest of growth might be due to accumulation of toxic waste products in the medium and the volume of the culture fluid supplied was increased to 100 c.c. in one pair of cultures in order to dilute the concentration of such by-products.

The same batch of medium was, of course, always used for experimental flask and control, equal volumes being measured into each at the same time.

In the second series of cultures (Series B) the culture medium contained, besides the above constituents, small known amounts of combined nitrogen in the form of Witte’s peptone, potassium nitrate, or extract of Calluna root. The object of this series of experiments was to determine whether small quantities of fixed nitrogen had any effect on the nitrogen-fixing power of the organism.

In all cultures of this series, 100 c.c. of the culture fluid were subjected to analysis for nitrogen before and after growth. The difference in nitrogen content gives the amount of nitrogen fixed by the fungus.

At the end of the growth period the culture flask was disconnected, the glass tubing dipping into the fluid was rinsed down with ammonia-free water, and the following determinations were made :

• Microscopic examination of the culture for the presence of foreign organisms.

• Change in pH value of the culture medium.

• Dry weight of the mycelium.

• Amount of glucose used during growth.

• Amount of nitrogen fixed.

Before any analyses of the culture were made, a small piece of the mycelium was withdrawn, mounted in gentian violet glycerine jelly and examined microscopically for fungal or bacterial contaminations.

The change in pH was measured by matching against standard buffer-indicator solutions before and after the experiment. At the end of the growth period, 10 c.c. of the culture fluid were withdrawn with a pipette, run into a test-tube, and the correct amount of indicator added. The results of these determinations will be found in Tables III and IV (pp. 178, 179).

Subsequently, the contents of the test-tube were replaced in the flask and the test-tube and pipette washed out with ammonia-free water, the washings being added to the rest of the liquid.

The culture fluid containing the mycelium was filtered through a dried and weighted ashless filter paper, the flask being thoroughly washed out with ammonia-free water and the washings being passed through the filter. After further washing the filter and mycelium were dried at 120 ° C. and weighed. The values obtained from these weighings will be found in Tables III and IV (pp. 178, 179). The filtrate, together with all washings, was received in a graduated 250 c.c. flask.

The filtrate was made up to 250 c.c. with ammonia-free water; 25 c.c. were withdrawn into a 100 c.c. flask and distilled water added to the graduation mark; this solution was used for the sugar estimations.

The analyses were carried out by Benedict’s method. The Benedict solution was restandardised before analysing each set of culture fluids.

The results of these estimations are shown in Tables III and IV (pp. 178,179).

Since we are here dealing with minute amounts of nitrogen (if any) in a large volume of fluid, it was considered that it would be preferable to use the whole of the culture fluid for one estimation of the macro-Kjeldahl type, rather than to divide the solution into aliquot parts and use a micro-Kjeldahl method. The only disadvantage of this procedure is that accuracy cannot be checked by carrying out duplicate analyses. Moreover, should any mischance occur during the final estimation, it involves waste of one month’s growth. Great care is thus needed during the estimation, and every possible precaution was taken to avoid the slightest errors.

The whole of the 225 c.c. of the solution remaining in the 250 c.c. flask, therefore, together with the dried mycelium and filter paper, were used for each nitrogen estimation, which was carried out by the Kjeldahl method modified to suit the particular requirements of the case, the chief difficulty being that we are here dealing with large quantities of sugar and very small quantities of nitrogen.

The diluted fluid, which represents nine-tenths of the culture fluid (one-tenth having been removed for sugar estimation) is placed in a 500 c.c. Kjeldahl flask and 30 c.c. of concentrated H₂SO₄ is gradually added. The whole is concentrated to small bulk. When charring sets in, the dried filter and mycelium are added to the contents of the flask; 10 drops of a saturated solution of CuSO₄ are also added by means of a pipette, together with about 3 gm. of solid K2SO4. The flask is transferred to a fume cupboard and into the neck is inserted a loosely fitting glass hood connected to a small flask which in turn is connected to a water pump. This arrangement very much lessens the inconvenience caused by the acid fumes, which are otherwise very objectionable owing to the high sugar content of the liquid, some of the acid gas condensing in the trap, the rest passing through to the water pump where it is dissolved and carried away with the pump water.

Great care in the regulation of the heat supply is necessary during the early stages of charring, otherwise there is a risk that the liquid will froth over the top of the flask neck. Later, frothing subsides and the flask may be safely left.

Owing to the large amount of sugar present the digestion takes about four hours, and is continued for 15 minutes after the liquid has turned from brown to bluegreen. The flask is then cooled and diluted to about 100 c.c. with ammonia-free water, the liquid decanted into a round-bottomed 500 c.c. distilling flask and the washings of the Kjeldahl flask added. Steam distillation was used because although slow it is the safest method. Ammonia-free water is boiled in a steam can for 10 minutes before the distillation is started. The receiving flask contains 25 c.c. N/15 H₂SO₄ and about 25 c.c. ammonia-free water. Forty per cent, sodium hydrate is made up with ammonia-free water and is run into the distilling flask by means of a tap funnel after all connections have been made. The solution is distilled for an hour and a quarter, after which the condenser tube is rinsed down within and without into the receiver with ammonia-free water. The contents of the receiver are now boiled for five minutes to expel any carbon dioxide which is sometimes present if the original soda contains any carbonate.

When the flask is cool its contents are titrated against N/15 NaOH. The indicator used was de Wesselow’s Indicator (a mixture of methyl red and methylene blue). This has a much clearer end point than methyl red alone, as the colour change is from purple to green instead of from pink to yellow. Throughout the series of experiments the same two burettes were used, one for acid and one for alkali. This obviates the necessity for calibrating the burettes.

After each analysis, a blank estimation was carried out using the same quantity of culture fluid and the same amounts of other materials (sulphuric acid, etc.), in order to determine whether any correction was necessary to allow for contamination by nitrogen compounds of any of the reagents used.

The results of the analyses of the two series of cultures, A and B, are shown in Tables III and IV (pp. 178, 179).

Below are given the details of a typical analysis of a pair of cultures (A 1 No. 4 and A 2 No. 4) to indicate the order of values obtained.

10 c.c. N/10 oxalic acid ≡ 14.90 c.c. NaOH (three identical values).

Therefore normality of alkali = 1.0067 N/15.

25 c.c. H₂SO₄ ≡ 26.90 c.c. alkali (three identical values).

Therefore normality of acid = 1.0832 N/15.

After Kjeldahl distillation, 25 c.c. of acid (1.0832 N/15) ≡ 26.81 c.c. of alkali (1.0067 N/15), i.e. 27-08 c.c. N/15 acid = 26.97 c.c. N/15 alkali.

Therefore amount of N/15 acid already neutralised = 0.11 c.c.

Therefore amount of nitrogen present = 0.00010 gm.

After Kjeldahl distillation, 25 c.c. acid (1.0832 N/15) ≡ 26.30 c.c. alkali (1.0067 N/15), i.e. 27.08 c.c. N/15 acid ≡ 26.48 c.c. N/15 alkali.

Therefore amount of N/15 acid already neutralised = 0.60 c.c.

Therefore amount of nitrogen present = 0.00056 gm.

Subtracting blank nitrogen present = 0.00046 gm.

Multiplying by ten-ninths to correct for amount removed for sugar estimation, therefore amount of nitrogen fixed by fungus = 0.00051 gm.

After Kjeldahl distillation, 25 c.c. acid (1.0832 N/15) = 26.65 c.c. alkali (1.0067 N/15), i.e. 27-08 c.c. N/15 acid = 26-83 c.c. N/15 alkali.

Therefore amount of N/15 acid already neutralised = 0.25 c.c.

Therefore amount of nitrogen present = 0.00023 gm.

Subtracting blank nitrogen present = 0-00013 gm.

Correcting as above for amount removed for sugar estimation, therefore amount of nitrogen fixed by fungus = 0.00015 gm.

The nitrogen analyses of the seeds and seedlings of Calluna vulgaris for Section A of the present paper were carried out in a similar manner, except that in the absence of such large amounts of sugar the whole process was much more rapid.

To determine the accuracy of the method, Kjeldahl estimations were carried out on weighed amounts of asparagin dissolved in nitrogen-free distilled water and also in 7 per cent, glucose solution to ascertain whether the presence of large amounts of glucose affected the accuracy of the method.

The results yielded by five analyses of 25 c.c. portions of asparagin solution, either with or without glucose, were

From these five values the standard deviation and hence the probable error of the experiment was calculated. This gave a value forthe probable error of 0.00007 gm. on each determination. The probable error of the average is 0.00003 gm., which indicates that the method gives figures which are accurate to 0.0001 gm. or rather more.

In all cases but one, the cultures were found to be perfectly pure at the end of the growth period. The fungus forms a mat of hyaline mycelium which was entirely submerged in all cultures except those in which nitrate was supplied. These latter showed a good aerial growth. The mycelium was composed of sparsely-branched hyphae without pycnidia. Pycnidia were not formed in any of the cultures grown, no doubt owing to the fact that the strain used for inoculation had lost the power of pycnidia-formation from long cultivation outside the plant. Ternetz (1907) found that the amount of nitrogen-fixation in other species of Phoma was closely connected with spore formation and most of the nitrogen fixed was retained in the spores. Possibly the absence of sporing in our strain is correlated with the low values of nitrogen-fixation obtained.

Throughout the series of experiments no appreciable change in pH value was observed to take place during growth: ± 0.3pH unit was the maximum change noted and usually the final pH value of the culture fluid was not more than ± 0.1 pH unit different from its initial value.

This is of interest in view of the fact that in another series of experiments, carried out on the same fungus for a different purpose and not dealt with here, in which the ratio of nitrogenous to carbonaceous food material supplied was very much higher, the pH was found to rise very markedly during growth.

In Series A (no combined nitrogen), the average dry weight of fungus (9 cultures) at the end of the experimental period was 0-054 gm. in A 1 and 0-049 gm. in A 2. There is thus, on the average, a slightly greater gain in weight in the case of the cultures grown in the presence of gaseous nitrogen. This difference is of little significance, however, in view of the considerable fluctuation in value of dry weight when individual cases are compared.

When these deviations are examined statistically it is found that the difference between the mean dry weights of the two series is not greater than the standard deviation of the two series. Consequently the apparent difference is meaningless.

Individual variations in dry weight in either series of cultures may arise from several causes. Thus, the cultures are not all strictly comparable, e.g. No. 4 was grown for 14 days only, while No. 7 was grown in double the amount of culture fluid. Then the size of the inoculum, which cannot be perfectly standardised, exerts a marked effect on the amount of growth. In addition, there are variations in temperature to be taken into account. Although each pair of flasks was kept at the same temperature by immersion in a water bath, no attempt was made to keep the separate pairs at an absolutely constant or identical temperature. There could have been little difference in temperature between pairs of cultures grown simultaneously, as the water baths were on opposite sides of the same bench. Even in cultures grown at different times, the temperature variation never exceeded 3° C., but such differences might account in part for the variations in dry weight observed.

It is noteworthy that No. 8 and No. 9, which are anomalous with regard to nitrogen-fixation, show perfectly normal dry weights.

As judged by the general appearance of the mycelium in the culture flasks, no appreciable growth took place after the first two weeks. In confirmation of this general impression, one culture and its control withdrawn after a fortnight showed as large a dry weight and nitrogen content as the cultures grown for the longer period. As mentioned earlier (p. 176), it was thought that the observed arrest in growth might be due to the accumulation of toxic waste products and with a view to obtaining information on this point the volume of culture fluid supplied was increased from 50 c.c. to 100 c.c. in a pair of cultures (A 1 No. 7 and A 2 No. 7) in order that such toxic products might be diluted to one half the concentration.

Reference to Table III shows that A 1 No. 7 gave an unusually high dry weight and consumed an unusually large amount of glucose compared with the other 50 c.c. cultures, a fact consistent with the view that arrest of growth is caused by the production of toxic by-products. On the other hand, A 2 No. 7 shows no departure from the normal values for dry weight and glucose consumption, thus lending no support to this hypothesis, unless it can be supposed that absence of nitrogen gas is the limiting factor to further growth. (It should be noted that A 1 No. 7 and A 2 No. 7 is the one pair of cultures which gave any indication of contamination by a foreign organism.)

In any case it is perhaps rash to lay much stress on the results of single pairs of cultures, although it seems likely that there is an arrest of growth after the first two weeks which cannot be attributed to change in pH value or to lack of carbohydrate food material. Further work is obviously required before the cause of this arrest of growth can be definitely assigned to any one cause.

It is recorded below that addition of minute amounts of combined nitrogen produces a marked increase in the dry weight of the cultures, and there is no doubt that, even assuming the capacity for nitrogen-fixation in the fungus, the absence of a readily available source of combined nitrogen causes a check on the rate of growth. At present no data are available from Series B as to whether the duration of growth is also affected, i.e. as to whether the arrest of growth of cultures in Series A is caused by nitrogen starvation or toxicity of the culture medium.

To Series B (with combined nitrogen), the same remarks apply: there is a slightly greater dry weight in B 1 (average = 0.3134 gm.) than in B 2 (average = 0.2961 gm.) to which differential result, however, too much significance must not be attached for the same reason. Here allowance must be made when comparing the different pairs of cultures with one another for the additional variable factor introduced by variation in the kind and amount of combined nitrogen added.

It will be seen that the addition of even so small an amount as 0.0016 gm. of combined nitrogen to the 100 c.c. of solution has a most marked effect on the rate of growth of the fungus, the average increase in dry weight in Series B being more than six times that in Series A.

Attempts to determine the amounts of glucose used during the growth of the different cultures resulted in the obtaining of irregular values which appeared to bear no relationship either to the amount of dry weight or to the amount of nitrogenfixation. A large part of the irregularity in glucose values is probably to be traced to the difficulty of estimating with the required accuracy small differences in glucose content in a medium very rich in glucose.

In addition, the same factors referred to as favouring variations in dry weight will be operative here also.

Nevertheless, even when due allowance is made for these inaccuracies and disturbing factors the figures show that, on the average, cultures grown in the presence of atmospheric nitrogen use about 1.5 times the amount of glucose used by cultures grown without gaseous nitrogen. When considered statistically, the individual deviations in the amounts of glucose used could only account for about one-eighth of the difference which exists between the average consumption of glucose by cultures in Series A 1 and A 2, so that this difference must be of real significance and not only apparent. The calculation of standard deviation for the glucose consumption was performed on strictly comparable cultures only, i.e. on the cultures A 1 and A 2 Nos. 2, 3, 5, 6, 8, 9.

The consumption of glucose is nearly doubled when small amounts of combined nitrogen are added to the culture medium.

In Series A the results of nine pairs of cultures are available. No. 8 and No. 9 are somewhat anomalous and will be considered separately; regarding the other seven the following remarks may be made :

1. The amount of nitrogen-fixation is never greater than the 4th place of decimals in grams. (The probable error of the average value for nitrogen-fixation under the conditions used was calculated to be ± 0.00005 gm.) The amounts of nitrogen found to be present in Series A 1 are generally slightly above the experimental error: the amounts in Series A 2 are generally below or hardly above the experimental error.

2. The variation in amounts of nitrogen as between the different cultures may be real or apparent. If apparent, it may be due to absorption of ammonia during the Kjeldahl estimation. This error, however, should be eliminated by the blank estimations done on the same amount of medium before growth. These blanks also eliminate any error due to impurities in the medium itself. If real, the variation may be due to one or more of the causes mentioned in a previous section (p. 182), viz. differences in period of growth volume of culture fluid, temperature, and size of inoculum. The last of these is probably far too slight to cause appreciable difference in nitrogen content directly, but may do so indirectly in that it influences the amount of growth and hence possibly that of nitrogen-fixation.

3. With the exception of the pair of cultures A 1 No. 3 and A 2 No. 3, the amount of nitrogen found at the end of the experimental growth period was always slightly greater in the culture provided with a supply of molecular nitrogen; while the average for the seven pairs, including A 1 No. 3 and A 2 No. 3, shows nearly twice the nitrogen content when molecular nitrogen has been available.

4. The two pairs A 1 No. 8—A 2 No. 8 and A 1 No. 9—A 2 No. 9 are exceptional in showing a nitrogen content in Series A 1 of more than six times that of their controls in Series A 2, or of the average of the other seven cultures in Series A 2. This divergence from the normal is so great that the results from these two cultures cannot be accepted without a very thorough scrutiny of any possible sources of error. Such a likely source of error is not easy to find.

The high values of nitrogen obtained for cultures A 1 No. 9 cannot be due to impurities in the culture medium, since the same medium was used in the controls (A 2 No. 8 and A 2 No. 9).

Now the two pairs 8 and 9 were grown simultaneously and throughout growth A 1 No. 8 and A 1 No. 9 were coupled together by means of their gas supply. The taps admitting the gas supply to A 1 No. 8 and A 1 No. 9 were never open simultaneously; nevertheless indirect gaseous connection was established several times during growth while admitting oxygen to the culture flasks from the voltameter.

It might be suggested, therefore, that by some chance the voltameter gas had become contaminated with NH₃ or some other nitrogenous compound which had been utilised by the fungus in A 1 No. 8 and A 1 No. 9. A 2 No. 8 and A 2 No. 9 were supplied from a different voltameter and hence would not show high nitrogen values.

The inherent improbability of such an explanation, however, is strengthened by the following considerations :

Firstly, the electrolytic gas was obtained by electrolysis of a solution of specially pure Ba(OH)₂ between platinum electrodes. The voltameter liquid was kept as much as possible from contact with the air, nor would a concentrated baryta, solution be likely to absorb appreciable quantities of NH₃ from the air.

Then again, if the source of error lay in the voltameter gas, the other cultures in Series A 1 should also show high nitrogen values, as these were run from the same voltameter, although if the result were due to some temporary cause it might not show in the other cultures which were not grown simultaneously with A 1 No. 8 and A 1 No. 9.

Nor does it seem likely that the results can be traced to gaseous compounds of nitrogen in the sterilised air supplied to the cultures in Series A 1, since this air was passed through a heated copper tube and then bubbled, first through concentrated H₂SO₄, and then through concentrated NaOH ; moreover, the same apparatus for supplying sterile air was used for other cultures in Series A 1, though not for all.

It may be that the conditions under which these two cultures were grown were, in some unrecognised way, specially favourable to nitrogen-fixation (e.g. traces of some organic compound which stimulated fixation may have been present in the medium).

It is a curious feature of these two cultures that neither the dry weight nor the sugar consumption was abnormally high. That there is no necessary correlation, however, between the amount of sugar consumed and the amount of nitrogen fixed is indicated by the results given with different organisms such as Clostridium, Azotobacter and Phoma (Table I, p. 168).

(5) In Series A 2, the nitrogen found on estimation obviously cannot be due to fixation of atmospheric nitrogen since none is supplied. It may be due to that introduced with the inoculum, or to ammonia absorption from the air during analysis. Neither of these are likely sources of large error, and in point of fact the amounts of nitrogen found were in almost all cases very small.

A statistical analysis of the results obtained from Series A 1 and A 2 shows that if the cultures 8 and 9 are included, the excess of nitrogen-fixation in Series A 1, over that in Series A 2 is five times greater than it could have been, supposing the differences to be due merely to individual variations such as always occur in members of the same series.

Thus if we include these two cultures we may conclude that Phoma radicis Callunae possesses considerable power of nitrogen-fixation.

But since values comparable to those obtained in the above two cultures could not be obtained in all the cultures it is scarcely safe to base conclusions on the results of these two.

If cultures 8 and 9 are ignored, the excess of nitrogen fixed by cultures in Series A 1 over that fixed by cultures in Series A 2 is still slightly higher than could have been the case if this apparent excess were in reality due to individual variations of the kind mentioned above.

These results, calculated on the basis of standard deviations, show that the “fixation values” though small are of real significance and not only apparent.

Although in these experiments, with the exception of 8 and 9, the degree of nitrogen-fixation proved to be slight, it is possible that fixation might be more vigorous under other conditions, as for instance an altered sugar or phosphate content, etc., of the medium. The medium employed was that used by Ternetz, but her reasons for supposing that it supplies conditions most favourable to nitrogenfixation do not appear to us conclusive. Added to this, the fungus culture available had been isolated from the plant more than two years previously and had undoubtedly lost, in addition to its power of pycnidia formation, part of its physiological vigour as was evidenced by its feeble growth compared with the vigour of the freshly extracted mycelium. It may well be, also, that nitrogen-fixation occurs more vigorously when the endophyte is growing in conjunction with the Calluna plant, where the products of fixation may be removed as fast as they are formed, than when the fungus is growing under the unnatural conditions of the experiment.

For these reasons it is felt that it would be unjustifiable to assume that because the degree of nitrogen-fixation was small in the above experiments this capacity for using molecular nitrogen is not of considerable importance to the plant growing in its natural environment.

Series B. It was thought that minimal quantities of combined nitrogen might stimulate fixation. Series B represents the results obtained when known small quantities of nitrogenous compounds in the form of peptone, nitrate and extract of Calluna root were added to the culture medium.

In Series B, there is also on an average an excess of nitrogen-fixation in Series B i, over that in Series B 2 and calculations of standard deviations show that this excess is four times too great to be due merely to variations such as occur between individuals of the same series. Therefore we may conclude that when supplied with small quantities of combined nitrogen, the endophyte of Calluna continues to fix atmospheric nitrogen at approximately the same rate as when no combined nitrogen is supplied. It was not found that, in the concentrations used, the nitrogen compounds supplied had any accelerating effect on nitrogen-fixation.

Zoond (1926), however, found that nitrogenous compounds (nitrates, peptone, tyrosine, plant extracts) did affect the nitrogen-fixing capacity in Azotobacter, the nitrogen fixed increasing with the nitrogen supplied up to a maximum after which increase in the nitrogen supplied caused a falling off in fixation.

Hills (1918) working with Azotobacter and with Bacterium radicicola found that potassium nitrate increased the nitrogen-fixing capacity, proportionally to the concentration of nitrate supplied.

It is therefore possible that the same phenomenon occurs with Phoma also, but that we have not used concentrations of nitrogenous compounds suitable for its demonstration.

The Calluna seedlings and pure culture of the endophyte were kindly supplied by Dr M. C. Rayner, and we have to acknowledge gratefully the very valuable help we have received from her in the course of these investigations.

We wish also to express our indebtedness to the Council of Bedford College for the award of an Anonymous Studentship which enabled one of us to carry on this work.

The time required to clear the apparatus entirely of original gases by the action of the pump and replace them with the electrolytic gases may be determined in two ways : by actual measurement and by calculation.

1. The apparatus is filled with CO₂ from a CO₂ cylinder by passing this gas through for half an hour. The end point, when the apparatus is full of CO₂, can be determined by collecting the issuing gas in a burette of acidulated water, running in NaOH, and noticing that there is complete absorption. [The CO₂ cylinder is attached at X (Fig. 1, p. 172), tap F being open and E closed; the NaOH is, of course, omitted from the U-tube B.] A wash-bottle containing NaOH is now attached at X in place of the CO₂ cylinder so that when the pump is set going CO₂-free air will be sucked through the apparatus. A couple of weighed potash bulbs are attached to the exit tube at F and the pump started. The potash bulbs are detached and weighed at intervals and indicate, by ceasing to gain in weight, when all the CO₂ has been swept out. This also gives a rough measure of the volume of the apparatus by calculating the volume of CO₂ from the increase in weight of the potash bulbs.

   Below are the figures obtained from two experiments:

   

   

   View Large

   Thus the apparatus takes about three hours to clear from CO₂. The average gain in weight is 0.5032 gm.; this weight of CO₂ at 17 ° C. and 761 mm. of mercury (the temperature and pressure conditions of the experiment) occupies a volume of 272 c.c.

2. The time taken for the air in the apparatus to be replaced by another gas can also be calculated theoretically.

The pump was timed and found to do on an average 1080 strokes per hour. By collecting under water and measuring the gas passed through in a given time, the amount sent through by each pump stroke was found to be approximately 1 c.c. Suppose we take 400 c.c. as the volume of the apparatus. (This volume is greater than that obtained in the previous experiment, but subsequently the 200 c.c. culture flasks were replaced by 250 c.c. flasks; moreover it is better that this estimate should be on the high side.) Hence we have:

Thus this second method also indicates that the amount of the original gas left in the apparatus after three hours is inappreciable ; and that after pumping the voltameter gases for six hours through the apparatus this may be regarded as quite free from nitrogen. Actually there is some sweeping action as well as dilution by the incoming gases, so that the original gas will be replaced more rapidly than indicated by the above figures.