dilute

More than 2,000 yearD ao ::enophon noted. the toxicity o honey m.de fro!1 the flowers of certain sreòie, o± Rhododendron. Dunn,: the past 5 rears 1nvetiation carried. on, both in EiroDe U1d in Japan, have resiüted in the isolation of toxic rriflcip]-e or rinci'1es from variotts Rho.odend.rons, wh1e no toxici!;y ws observed. in otheru. o investigation apnears to have been made on R. ca1ifornic, a1thouh it lias been reported. as rOISOnOUS to shieD. The resent invetiation was nade in n atterpt to isolate the oieonous anroed.otoxin fror the leaves o this plint.

as rOISOnOUS to shieD. The resent invetiation was nade in n atterpt to isolate the oieonous anroed.otoxin fror the leaves o this plint.
Of the several raetho&s of isol:tlon used, on1r two roduced results. These were modiÍications o Hard..kar's method of isolation from a ïater e:tract, ihich had. been prepared. after :llling witb boiling water any enzyme Tresent in the leaves. The two rnod.ifications were in the method. or uriuication of the a ieoiìs extract. The first consisted. in treatment with neutral lead acetate solution and. hydrogen sulfid.e cas. The second.
(Britt1 nethod.) ef±ected. Isolation by fixing vrith nagnesiurn oxide the tannins prosent in the rater extract, treating it with 95 alcohol, a:d concentrating to produce fine crystalline needles. The crude prod.ucts, when c-iven subcutaneously to experimental animals in aiiounts twice as g,eat as the mininn lethal d.ose of' and.romed.otoxin, were usually ot,nd. :o be i:iti'. Further anirrtal exïeriraentation is, however, necessary using larer rlose since the crude roduct ra:. contain lar{e e.rnounts ¿if irrrntrities and. conseTlently small arnoirnts of the poison. color iections or the substance obtaIned. from Britt's method.
'-:are similar to that oÍ androneclotoxin, but its melting point v:as conid.erably below that reported Íor this coaoirnd.. It is doubtful, thcreí'ore, whether and.romedotoxin has been isolated Irorn the leaves of R. cali'ornici,n, but the inc.l conclusion can not be dravrn at oresent. Pirther investiration nust be carried. on to mirify and. i&enti±y the crystals vrhieh vree IsOlatEd, and to learn more concerning their physiological action.   (6) in 1882 from Andromeda japonica and called by him, after the Japanese name of the plant, "Asebotoxin." Pluge (ib) diseoverd it about the same time, first in the above plant and later in Andromeda 2.ifolia (le), A. calyculata (19), A. catesboei, and in several other Ericaceae not belonging to the enus Andromeda, e.g., Azalea indica, and Rhododendron ponticum (2C).

Pluge gave it the name "Andromedotoxin."
A systematic investigation of the substance, chemical arid pharmacological, was made by H. De Zaayr (b) in 1886, a smary of which was published (in Dutch) (b) by Plugge (17) in 1887. An extensive pharrncological investigation was made by S. W. Hardikar (7) in 1922, however ver.r little is known about the chemistry of axdromedotoxin.
Plugge (16) has investigated a number of plants which belonto the ricaceae family and determinrd the presence or absence of andromedotoxin in them, but his list is, of courEe, far from complete. It has been reported from the Pacific coast that the leaves of Rhododendron californicwn Hook., which belogs to the Tricuceae family, 2 are poisonous to sheep. The present investigation was undertaken to determine whether or not these le&ves are highly toxic, and if so, to Lscertain whether this toxicltv is due to andromedotoxin.  (22)). Those in which the poison was not found are as follows: yro1a macuiata L.

CHMI CAL
Chemically andromedotoxin has not been well investigated and needs further study. ccording to Eykrnan (6) aseotoxin, which is said to be identical with ndromedotoxin is toxic g1ucosid; while Pet6rson (15) sttes that andromedotoxin is an indifferent non-nitro, enous organic comound and not a glucoside. The latter viw has been suoported by many other investigators since andromedotoxin after hydrolysis does not reduce Fehling's soLiti on.
TykmEn describes andromedotoxin or asebotoxin as a colorless substance, soft while moist, but brittle and transparent when dried at entle heat. When covered with water it cakes together, becomes soft below 100° C., but does not melt until the temperature l2.0 C. is reached, and then forms a trLnsparent light-brown mass. Ìirs. Stedman (6), who made a chemical investigtion of this coinpound recently under Professor Barger, states that the pure andromedotoxin crystallizes in fine, white needles containing no water of crystallization, and mnelts at 258° C. The rneltin point of this comoirnd iven in Beilstein L2) is, however, 228-229° C. and that by Chu and How ( 3.) There is some disagreement among investigators cmcerning the molecular weiht and chemical formula for this compound.
Eykman (6) mentions another color reaction of the corapound with strong hydrochloric acid. Vihen strong hydrochloric acid is added to its alcoholic solution, a magnificent blue color develops, accompanied by a peculiar odor resemblin: that of Spiraea ulmaria. This blue color, under various conditioi, changes t: violet-red to reddish-gray.

PHARCOLOGICAL
Sever1 epidemic poisonings of sheep from eating rhododendron have b';en reported by Parkinson (13) and b Slipper (25). The chief symptoms are stated to be salivation, vomiting, weakness, staggering and, in a few cases, death.
The lethal dose of androraedotoxin by subcutaneous administration varies in different animals. In the case of the frog it appears to lie between 2.5 and rngm. per kilo. of body weight, but the rabbit requires only 0.3 rngrn.
1.4 rrìgm. androiedotoxin per kilo. hypodermically caused death in 15 minutes; while .28 rngin. per kilo. hypodermically cailsed death in 12-24 hours in rabbit. De Zaayer (5) likewise has shown that andromedotoxin is less toxic to lower forms of life than to the higher animals. There is no iriflìence upon unicellular organisms 1ie ?aramoecium Coli and Opima ranarum even in a satirated solution. Lumbricus terrestris remained quite normai in a solution of 1:8000. In a strength of 1:400, the worms at the commencement made active movements, became quiet shorti afterwards, but death did not occur until after the laose of 24 hours. In frocs De Zaayer (5) observed vomiting, arrest of respiration, and a paralysis resembling that of curare. In mammals he decribes the occurrence of emesis and purging, severe respiratory disturbances and dysDnoea, convulsions, and death through arrest of respiration.
There is some difference in opinion as to the cause of the various symptoms which occurred upon administering andromedotoxin to the e:oerimental animals.
De Zaayer (5) , and Chu and How (4) agree as to the cause of respiratory depression. They think it is due to a direct action upon the respiratory center because it appears in the frog before any symptoms of general paralysis, and in rabbits the nerve ends are shown to be intact by the convulsions which attend the failure of respiration and which may be prevented by artificial inflation.
The respiration in the rabbit is retarded without any previous acceleration; the chances observed could not be due to an action on the vagus endins in the lunas, for reaction of the vagus does not produce similar changes. This VicW is supported by Hayashi and ruto (e), Archangeisky (1) and payne (14). Hardikar, how6ver, attributes this depression to an action of the poison on the vagus, for the poison t first stimu-1 tes and then par1yzes the termintions of the vagus.
ccording to Ch. and How (4) increase of blood pressure by andrornedotoxin is due to a direct action on the heart, but Hardikar (7) states that it is due to an effect upon the vagus. Although Hardikar was unable to determine the seat of emetic action, De Zaayer (b) states that it may be caused by a direct acion of the sbstance itpon the vomitin center, since vomiting occurred in a dog three minutes after a subcut&neous injection. All the previos investigators, however, agree to the cause of death. In small doses, death occurs from the arrest of respiration from paralysis of phrenics. With a very 1are dose, acording to Hardikar ('r), death is due to a direct action upon the heart, the vntric1es beine arrested in diastole or partial systole.
In addition to those mentioned above, the following oharinacological actions of andromedotoxin hve been observed by }Jardikar (7' 1. D"spnoea of the asthnitic type due partly to stimulation of afferent fibers and pertly to spasm of the bronchial muscle from stimulation of its motor nerve, the vagus.
2. Increse o bronchial secretion. 3. Slowing of the heart and fell of blood pressu.L'e followed by cee1erat1on and rise of pressare. 4. Repeated evacuation of the bowels.
(He attributes these four symptons to stimalation followed by paralysis of teninations of the vague).
b. Paralysis of the motor nerve ends in striped micle. While the paralysis is developin;, the muscle and nerve are more easily f ti:ued, but regain their excitability after a period of rest. In a stronger concentration, the poison also afc.cts the muscle substance itself which thus permanently loses its excitability.
The manifestation of this fatigue is seen beet and earliest in the nerves and muscles which hve to be constantly in action, viz, the phrenics and diaphrgrn.
The sane action has been observed by Hayashi and uto ( 8 6. A narcotic action upon the higher center in the brain, the spinal cord being not affected.

7.
A condition of arrhythmia in the heart dependent either upon a direct depressant action upon the conductivity tissue between the auricle and ventricle 1eding to heart-bloc:, or upon the exc1tb11ity of the ventricle itself.
The period required for diastolic relaxation of the ventricle is increased and the diastole is incomplete. 8. The perfused frog heart was arrested with the ventricle as well as auricle in diastole; while in frogs injected with a fatal dose, the ventricle ws arrested in tota' or partial systole with the auricle distended.
The perfused mammalian heart was arrested in systole of the ventricle, but in the death of an animal injected with the poison, the right side was distended &rid left empty or in partial systole. 9. Involintary muscle which is not supplied by the vagus was not affected. Thus it is seen that in sample "C", the weight of the air-dried leve: represented 42 of the weight of the fresh leaves.

Moisture and Ash con4erit
For the determination of the ioisture and ash con-tent of the sir-dried materials, the sample 1TA was used.
In determining the rìoistu.re content, the 1eves and stems were dried at 110° C. in an oven; and in determining ash c ntent the air-dried samole was incinerated in a crucible  (7), his methods of isolation were firt employed.

Method i -Alcoholic extract
The extract "F" was taken up in a small Q.uantity of water, precipitated with neutral lead acetate solution (ic), filtered, made free from the excess lead by treating with hydrogen sulfide gas and filtering.
The aqeous Nltrate thus obtained was evaporated to dryness on a waterbath.
The residue was dissolved in a small Quantity of 95c3 alcohol and precipitated with an equal volrne of ether.
This was repeated several times till the alcoh:1ether treatmert extracted nothing more.
At first a white milky solution was obtained, bit on staridin the solution became clear, leaving a heavy sticky brown precipitate at the bottor of t-ie container.
Hardikar (7) obtained from this alcohol-ether mixture a crystalline substance which was found to be ohysiDlogically inert.
There not being enough precipitte to colLect, the ether .s driven off and the clear alcoholic solution, "(", was kept for further experiment. Method 2 -Water e:<tract (Hardikar) One hundred grams of leaves of the sample "A", were moistened with boi1in water in a lar .e bker, which was then covertd tightly with aper, and allowed to stand for 15 minutes. The moistened drue boiled for twelve hours, replacing water lost by evaporation.
The water extract thus obtained was evaporated to a syrap consiste'cy, mixed with saw-dust which had previously been purified with chloroform, and then extracted with chloroform in a Soxhiet extraction apparatus.
The extraction was discontinued after being run for eighty hours although evaporation a small portion of the fresh extract still let sorne residue on a watch glass. The chloroform extract was evaporated to dryness, the residue left was purified by treating it several times with ether, which was decanted oÏf each time.
This prified residue was dissolved in n-butl alcohol to crystallize since Mrs. Stedman (6) su:ests that n-butyl alcohol is a good crystallizing solvent for andromedotoxiri.
Several crystalllzat.ions from n-butyl alcohol produced a yellow owder, "H", instead of the crystals which Hardikar was able to obtain. The tota weight of the yellow powder was 0.0842 gram, corresponding to a yield of 0.0842% based on air-dried drug.
This product was tested for toxicity on a 900-Gin.
guinea pig, using a dose of 2 mpm. in a half cc of physiological salt solution and was found to be inert.
The product did not reduce Fehli:'s so1utior either berore or after hydrolysis.

Method 3
If andromedotoxin is a glucoside as stated by Eykman (6, an enzyme will probably accomDany it in the plant, an the enzyme will hydrolyze the 1ucoside when 1eves are treated as in the previous experiments. ased on thiE assumption, it would be necessary, before attemptin to isolate the ßlucoside, to kill the enzyme, which might have caused the unsuccessful results obtained in the previous experiments.
Into a 12-liter balloon flash ws put six liters of' water, which w.s brought to boiling and leves were introduced while the water was boiling,.
This treatment i1ls any enzyme which might be present j_n the leaves. 3oi1inc wa continued for one hour, whereuDon the water was poured off and replaced with fresh water. In this manner, several batches of extract were made.
The combined extract was evaporeted to a small volurie on a water-bath. Half of this concentrated e:tra3t was mixed with saw-dust (previously purified with chloroform) and extracted with chloroform for thirty hours.
The chl3rofor:1 extract was evapzrated to a small volume and treated several times with ether, which ws decanted off. 2he purified residue thu.s obtained ws dissolved in n-butyl alcohol to crystallize.
A brown sticky mass, "I", with a sweetish odor was obtained.
in attemnt was made to remove the brown clor by recrystallizing it from n-butyl alcohol, but it did not chance the aDpearance of the substance.  (15) was tried, using 89 grains of the extract "D".
The extract was dissolved in a small amount of water, and extracted with a large q»uantity of ether to remove ether soluble impurities, especially chlorophyll, 'J". The aqueous portion was then extracted with chloroform in a seDaratory funnel and evaporated to dryness in air.
The result was very poor--nothing but a trace of chioroDhyll ws left in the evaporating dish. Etùr extract, TTJTT, and the ueous solution, "K", were kept for further investigation.

Method
According to Peterson (iö) andrornedotoxin is present only 0.005% in p1ans. It was, therefore, necessry to do the isolation on a larger scale than that had been done in the 'revious exrriments. In Beilstein (2) it is mentioned that the poison is rauch more soluble in cold than hot wL.t water extract wa than straining it been done.
Water placed and 1000 grams of er.
In this experiment, therefore, the strained off after being cooled, rather right after boi1in, which had revious1y in a large kettle s bro.ht to boiling the sample of leaves, "C", was intro-duced all at once. Boiling waF continued for thirty minutes then the mixtu2e was left to cool. The water extract was decanted off and evaporated to a syrupy consistency on a water-bath. s the second batch of the water extract was destroyed by tire, the third batch wa prepared and combined with the first.
The total extract, !TLTI, obtained was 519.0 grams with a percentage yield of 51.9o, based on air-dried leaves.
IIethod 5j: The bove water extract contained 1axe amounts of tannins, which had to be removed before proceedin-.
The following method for their removal was suggeEted by r. Britt, an assistant professor of pharmaceutical analysis in School of Pharmacy at Oregon State College.
Of the extract flLTT above, 29 grams were mixed with zoo grams of heavy magnesium oxide and niade into a paste, allowing a reaction to take place between the tannic acids and ma:nesium oxide to form an alcohol- An attempt to recrystallize this substance from n-butyl alcohol was not successful since the sticky crystals did not go into solution. The crystals, "N", were recrystallized from water, (N'), and rrom n-butyl alcohol (NTT), and dried in a desicator. The results compared with that of andromedotoxin are as follows:

4.
Toxieit of the crystals, "M" and "N" Into a guinea pig, weighing 672 grams, were injected 0.373 ingm. of the crystals "M" dissolved in a half c.c. physiological salt solution. The injection was followed by no observable physiological effects.
Into another guinea pig, weIghing 806 grams were injected 4.5 zngm. of the crystals "N" dissolved In a half c.c. physiological salt solution. This Injection also showed no physiological action. from leves of hododendron californicu.in. tremendous amount of time ws required to try each method of isolation, and with the limited time available it was not possible to cover the entire problem. Luch has been left untouched at present but the investiation will be continued to determine if possible Wiat really is the toxic constituent of the leaves of Rhododendron californicum.
If this be androm8dotoxin, then an explanation must be found for the anomaliøs in the melting point and in the action observed on guinea pigs.