Volatile flavor components of celery stalks (Apium graveolens VAR. dulce) as related to temperature and time in storage ...

MISSING IMAGE

Material Information

Title:
Volatile flavor components of celery stalks (Apium graveolens VAR. dulce) as related to temperature and time in storage - with further investigations on component distribution within the stalk
Physical Description:
95 leaves : ill. ; 28 cm.
Language:
English
Creator:
Ezell, Danny Odell, 1941-
Publication Date:

Subjects

Subjects / Keywords:
Celery   ( lcsh )
Vegetable Crops thesis Ph. D
Dissertations, Academic -- Vegetable Crops -- UF
Genre:
bibliography   ( marcgt )
non-fiction   ( marcgt )

Notes

Thesis:
Thesis--University of Florida, 1968.
Bibliography:
Includes bibliographical references (leaves 91-94).
Statement of Responsibility:
by Danny Odell Ezell.
General Note:
Typescript.
General Note:
Vita.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 000414257
oclc - 37526033
notis - ACG1417
sobekcm - AA00004940_00001
System ID:
AA00004940:00001

Full Text












VOLATILE FLAVOR COMPONENTS OF
CELERY STALKS (Apium graveolens VAR.
dulce) AS RELATED TO
TEMPERATURE AND TIME IN STORAGE-
WITH FURTHER INVESTIGATIONS ON
COMPONENT DISTRIBUTION WITHIN THE
STALK



By
DANNY ODELL EZELL











A DISSERTATION PRESENTED TO THE GRADUATE COUNCIL OF
THE UNIVERSITY OF FLORIDA
IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE
DEGREE OF DOCTOR OF PHILOSOPHY









UNIVERSITY OF FLORIDA
1968













ACKNOWLEDGEMENTS


The author wishes to express sincere appreciation to

the Chairman of the Supervisory Committee, Dr. D. D. Gull,

for his guiding assistance in these experiments. Appreci-

ation is expressed to Dr. R. C. Smith, Dr. B. D. Thompson,

and Mr. R. K. Showalter, who have contributed to his academic

training, direction of the research, and preparation of this

dissertation.

Special recognition is given to his wife, Wanda, who

has given understanding, encouragement, and long hours as a

dedicated typist to make this graduate study possible.















TABLE OF CONTENTS


ACKNOWLEDGEMENT

LIST OF TABLES


LIST OF FIGURES. .

INTRODUCTION .

REVIEW OF LITERATURE .

MATERIALS AND METHODS. .

Storage. .

Phase I. .

Phase II .

Phase III. .

Component Distribution .

Analytical Procedures. .

Organoleptic .

Extraction .

Chromatographic .

RESULTS AND DISCUSSION .

Organoleptic Evaluation.

Storage. .

Phase I. .

Phase II .

Phase III. .


Page

4 4


.... vii

* 1

* 3

* 9

. 9

* 9

S. 10






.... 10

... 10

S. 12



.. 16

..... 25

. 26

. 26

. 35


iii


c0


. v


* 0 0 a 0 0 0 0 .







Page

Component Distribution ........... 81

SUITIARY AND CONCLUSIONS . 88

BIBLIOGRAPHY . . 91













LIST OF TABLES


Table Page

1 Storage treatments used in Phase III. 11

2 Oven temperature and time relationships
for the matrix sequence used in all
chromatographic measurements. ... 15

3 Retention times on Apiezon L column and
average peak heights of peaks with re-
tention time greater than 90 minutes. 20

4 Odors detected at the exhaust port of an
Apiezon L column after injection of
1 ul of celery solvent extract residue. 21

5 The effect of low temperature (380F)
storage on the ratio of high/low
boiling fraction in celery flavor
extracts. . . 27

6 The effect of storage at 380F for 2 or 4
weeks upon the relative per cent of 9
major chromatogram components . 28

7 The effect of storage at 70 F for 5 days
upon the relative per cent of 9 major
chromatogram components. . 36

8 The effect of storage under market sim-
ulated conditions upon the relative
per cent of 9 major chromatogram
components. .. . 44. 44

9 Observed retention times of volatile
components of celery head-space
samples, using 2 column stationary
phases. ....... ............ 50

10 Relative per cent of total peak area
for the 9 peaks measured in head-
space analyses.. .. .. 51










11 Mean peak area (fresh weight basis)
for each of the 9 peaks analyzed
in head-space measurements as re-
lated to storage temperature and
duration ... .. .. ... 52

12 Per cent dry weight of celery as
related to storage temperature
and duration ....... ....... 65

13 Mean peak area (dry weight basis)
for each of the 9 peaks analyzed
in head-space measurements as re-
lated to storage temperature and
duration . . 66

14 Ratios of high/low boiling fractions
from chromatograms of extracts pre-
pared from various parts of the
celery plant .. . 81


Table


Page













LIST OF FIGURES


Figure Page

1 Chromatograms prepared from (A) dry
ice-acetone and (B) liquid nitrogen
extracts. . . 17

2 Chromatogram prepared from aqueous
extract as described in extraction
procedures. . . 18

3 Infrared spectrum of compound
believed to be n-butyl phthalide. 23

4 Chromatograms from celery extracts
prepared (A) after 10 weeks' storage
at 380F and (B) at harvest. . 30

5 Mean ratio of high/low boiling com-
ponents as affected by storage for
5 days at 70F (difference signif-
icant at 0.05 level). . 32

6 Mean peak area of chromatograms pre-
pared from celery extracts taken at
harvest and after 5 days' storage at
70F (difference significant at 0.01
level). . . 33

7 Chromatograms from celery prepared (A)
after 5 days'storage at 70F and (B)
at harvest. . ...... 34

8 Mean ratio of high/low boiling com-
ponents of the chromatograms from
the control and market simulation
treatment (difference significant
at 0.05 level). . ....... 38

9 Mean peak area of low and high boiling
fractions of the chromatograms of
the control and market simulation
treatment . 39


vii







Figure


10 Mean peak area of d-limonene and
n-butyl phthalide for the control
and market simulation treatment
(""'change in peak area significant
at 0.01 level). ............ 40

11 Mean peak area of all peaks in the
low and high boiling range ex-
cluding limonene and butyl
phthalide, respectively ......... 42

12 Chromatograms of extracts prepared
from (A) celery stored under
market simulated conditions and
(B) freshly harvested celery. 45

13 Chromatogram from head-space mea-
surements of celery volatiles with
peak numbers according to reten-
tion time (Apiezon L) .. . 47

14 Chromatogram obtained when celery
head-space volatiles were chro-
matographed on Carbowax 20 M. 49

15 Trends in the change of peak area
(fresh weight basis) as a result
of duration in storage at 38 F. 57

16 Peak area (fresh weight basis) of
celery samples after 2 and 4 weeks'
storage at 380F and subsequent
storage after each for 1 day at
70F and 8 days at 50F (including
all peaks). ............. 59

17 Peak area (fresh weight basis) of
celery samples at harvest, after
storage at 38? for 2 and 4 weeks,
and after storage at 450F for 2 and
4 weeks for each of the 9 peaks
measured in head-space analyses 62

18 General trends in total peak area
(fresh weight basis) changes
according to temperature and dura-
tion of storage at 450F, 380F, and
subsequent storage at 50F and 70F
after storage at 380F . 63


viii


Page









19 Trends in the change of peak area
(dry weight basis) as a result of
duration in storage at 380F . 70

20 Peak area (dry weight basis) of
celery samples after 2 and 4
weeks' storage at 38F and sub-
sequent storage after each for
1 day at 700F and 8 days at 50F
(including all peaks) .. . 73

21 Peak area (dry weight basis) of
celery samples at harvest, after
storage at 380F for 2 and 4 weeks,
and after storage at 450F for 2
and 4 weeks for each of the 9 peaks
measured in head-space analyses 75.

22 General trends in total peak area
(dry weight basis) according to
temperature and duration of storage
at 40F, 380F, and subsequent storage
at 50F and 700F after storage at
38F. .................. 77

23 Retention time of 5, 6, and 7 carbon
straight chain alcohols for polar
and non-polar chromatographic phases. .. 83

24 Chromatogram prepared from an extract
of the top (leafy) portion of the
celery plant. .. ...... 84

25 Chromatograms prepared from celery
extracts of (A) the inner and (B)
outer portion of the stalk. . 85


Figure


Page













INTRODUCTION


With the increasing sophistication in analytical

techniques and instrumentation has come a greater knowledge

on the volatile constituents of foods. Although the flavor

of celery (Apium graveolens var. dulce) is quite distinctive,

it has received relatively little attention as compared to

other fruit and vegetable flavors.

Vegetables are in many cases quite difficult to

examine chemically. Flavor substances frequently occur in

concentrations of a few parts per million, while accompanied

by other organic materials and large amounts of water.

Several extraction (38) and analytical (39) techniques have

allowed increased resolution of these flavor substances and

a decrease in interference from foreign materials.

Celery is a perishable agricultural commodity; how-

ever, with proper handling it can be maintained in a market-

able condition for moderate periods of time. The long

storage durations involved in shipping Florida celery to

Europe have imposed more stringent demands on the keeping

quality of fresh celery. Quality reduction of celery due to

storage duration is usually manifest by water loss, pithiness,

loss of green color, decrease in sucrose and increasing

toughness. While flavor changes do exist as a result of





2

storage, it has not been determined if these flavor changes

are caused by changes in the volatile flavor constituents

or if these changes result in a detriment to quality.

This research was initiated to: i. select a proce-

dure for extraction and measurement of the volatile flavor

components in celery stalks; ii. observe whether differences

occur in the chromatograms of celery extracts as a result of

storage temperature and duration; iii. observe differences

in the chromatograms of head-space volatiles (vapor pressures

sufficient to permit analysis in the vapor phase without ex-

traction) obtained from celery as a result of storage temper-

ature and duration, iv. investigate the flavor profile of the

various portions of the stalk of celery.













REVIEW OF LITERATURE


The maintenance of quality of fresh celery in storage

involves several aspects of applied physiology. Corbett (8)

indicated that several physical and chemical changes were

apparent during the storage of celery. Chemical analyses

of different parts of the plant showed marked changes in

reducing and total sugars and also both soluble and insoluble

nitrogen. Reducing sugars decreased in the leaves from har-

vest to the end of storage. Young (45) and White-Stephens

(43) also concluded there was an increase in dry weight in

the crown portion of the celery plant with a corresponding

decrease in the outer leaf portion, and the outer leaves of

the pascal celery plant were of definite value in growth of

the heart petioles during storage. White-Stephens also

observed a marked increase in sucrose and polysaccharides in

the inner petioles coincident with a similar decrease in the

outer petioles. Corbett (8) concluded that soluble nitrogen

increased in both leaves and stalks from harvest to near the

end of the storage period,at which time there was a very

marked decrease in the leaves. This decrease of soluble

nitrogen in the leaves resulted in an increase in the inner

petioles.

Hall, et al. (24) cautioned against making quality





4

comparisons between different varieties of celery at differ-

ent maturities. Single varieties of celery seemed to follow

the same pattern of change in chlorophyll content, alcohol

insoluble solids, total sugars, crude fiber, and dry weight

with regard to increasing maturity. Three varieties followed

the same general trend, but varied in the time the lowest and

highest points were reached.

Organoleptic (taste) comparisons have been made be-

tween celery (Utah 52-70 variety) grown in Florida and the

same type celery produced in California. When considering

celery from an April harvest, the taste panel indicated the

Florida celery had a superior flavor, but was tougher and

had more fiber. However, when comparisons were made with

celery from a May harvest, the rankings indicated that the

California celery had a more desirable flavor and was less

bitter, but was tougher and more fibrous than the Florida-

grown celery. In these experiments it was noted that a low

potassium-sodium ratio was associated with a bland flavor.

The celery from the two areas was grown by commerical pro-

ducers, was the same age, and was handled under similar

conditions.

The price of Florida celery has frequently been below

that of celery produced in other areas. While these opinions

were not unanimous, the reasons given for the reduced price

included poor appearance, bitter flavor and toughness (22).

While making observations on the flavor of celery,

Hall (23) described a bitter fraction and a burning-numbing







sensation from celery. The bulk of the bitter flavor was

associated with the dark green outer layer of the petiole

and did not appear to be associated with the burning-numbing

sensation. Description of the various flavors were: salty

flavor, radish-like flavor, and a hydrocarbon-type flavor

described as kerosene-or turpentine-like. The panel occa-

sionally noted a sweet flavor. Hall also found a considerable

difference in the composition and organoleptic ratings of

outer, inner, and heart petioles of celery (21). The com-

position of these petioles was influenced by the temperature

at which the celery was stored. In addition to the dif-

ferences found between petioles according to position, there

were differences in flavor and composition between middle

and upper portions of the outer, inner, and heart petioles.

Using chromatographic analyses, Gold and Wilson (14)

showed that not all flavor constituents were in the juice

of celery stalks. The chromatograms prepared from juice and

puree showed common peaks, although the chromatograms were

not identical. An organoleptic study was performed on the

juice of celery from the top (leafy) and basal portion of

the celery stalk. The taste panel was able to differentiate

between the juice from the top (leafy) and basal portions

of stalks at the 0.01 level of significance. However, Gold

and Wilson were not able to differentiate between the chro-

matograms prepared from the two juices. There was no dif-

ference between chromatograms prepared from fresh and

frozen juice samples.







Pan (36) characterized the burning-numbing taste of

celery using gas-liquid chromatography. This chemical had

no unsaturated bonds or nitro groups. He found the presence.

of aldehydes, carbonyl, phenolic, hydroxy, and aromatic

groups but the fraction was not a lactone. In further re-

search (37), Pan described a new technique in the isolation

of a bitter principle from celery. This principle was

cationic at pH greater than 7, soluble in polar solvents,

and fluorescent under ultraviolet light. The bitter

principle could not be steam distilled from celery and was

located in the dark green portion of the petiole.

The complex mixture which constitutes the flavor and

odor of celery received considerable attention as early as

1897. Ciamician and Silber (7) reported the following

terpenes contributing to the flavor of celery: limonene,

myrcene, and an isomer of apiol. Of the lactone fractions

identified, sedanolide and sedanonic anhydride were pro-

posed as being of primary importance in the odor of celery

seed. Following the identification of sedanolide and

sedanonic anhydride by Ciamician and Silber, Berlingozzi and

Clone (4) undertook a study of the chemistry and odor

characteristics of alkyl and alkylidene phthalides. Working

with 6 -dihydrophthalide, 6 -tetrahydrophthalides, and

hexahydrophthalides, they found when one of the y-carbon

hydrogens was replaced by an alkyl group, a celery odor

was noted. When both were replaced by alkyl groups, the

odor was less intense. Celery odor was most intense when







the r-carbon hydrogen was replaced by the alkylidene group.

Intensity increased as carbons increased from 1 to 4.

Barton and DeVaries (2) analyzed celery oil and re-

ported the isolation of butyl phthalide instead of sedano-

lide when using Ciamician's method (7). They assumed that

an unstable sedanolide might be changed to butyl phthalide,

and proposed the structure of an ,, G-unsaturated lactone,

neocnidilide. Mitsuhashi and Muramatsu (33) further pro-

posed that sedanolide is at least a mixture of neocnidilide

and butyl phthalide.

Recent investigations concerning the flavor and odor

of celery were conducted by Gold and Wilson (14, 15, 16) and

Wilson (44)- In 1963 (16), they listed 38 compounds which

were identified as volatile components of celery. While

most of the compounds listed probably make some contribution

to the composite flavor and aroma of celery, 6 are of pri-

mary importance: 3-iso butylidene-and 3-iso validene-3a,

4-dihydrophthalide; 3-isobutylidene-and 3-isovalidene-

phthalide; cis-3-hexen-l-yl pyruvate; and diacetyl. The

4 phthalide derivitives were found in a ratio of 6:3:1:1.

The difficulty of separating the elements of a

chemical mixture into their respective flavor potencies has

been pointed out (17, 40). Guadagni, et al. (18) found

an additive effect of various chemicals even when the

chemicals were present in sub-threshold concentrations.

This difficulty was further emphasized in attempts to mea-

sure sensory responses for food products by direct injection







of aqueous vapors into a gas chromatograph (6, 29, 35).

Newar (35) noted that the concentration of a given compound

in a vapor phase at a given temperature is affected by:

vapor pressure of the compound, type of media in which it is

distributed, degree of solubility in the media, concentration

of compound in liquid phase, and its miscibility with other

organic compounds. Therefore, a decrease in liquid con-

centration does not necessarily mean a decrease in vapor

concentration.

While there has been much research concerned with

the quality of celery in storage and with the volatile flavor

components of celery, little research has been conducted to

investigate changes in these volatile components during

storage, and it is to this question that the following work

is addressed.













MATERIALS AND :rOTHODS


Florimart and Utah 52-70-2-13 (commonly known as

Florida-2-13) cultivars of celery were obtained from commer-

cial growers. Size Number 3 stalks (36/crate) were harvested

90 days after transplanting, packed, hydrocooled, and trans-

ported immediately to the laboratory. Preliminary investi-

gations were conducted to determine the most satisfactory

method for extraction of celery volatiles and measurement of

the various parameters.


Storage


Phase I

The initial study was established to determine the

effect of low and high temperature storage on the composition

of flavor extracts prepared from celery stalks. Low tempera-

ture effects were determined by comparing extracts from

freshly harvested Florimart celery with those from celery

stored 2 or 4 weeks at 38 F. High temperature effects were

determined by comparing extracts from freshly harvested

celery with those from celery stored 5 days at 70 F. A

randomized block design was used with 3 harvest dates con-

stituting blocks. A non-replicated study involved weekly

extracts of celery stored at 380F for 10 weeks.







Phase II

Florimart cultivar was also used in a storage treat-

ment to simulate market conditions. Treatment involved

placing the stalks at 450F for 2 weeks and transferring

samples to 50 F for an additional week. The experiment was

a randomized block with 2 harvest dates as blocks and a

duplicate analysis of each sample.


Phase III

Florida 2-13 cultivar was used for aroma (head-space)

measurements. Treatments employed for these analyses are

listed in Table 1. Statistical design of the experiment was

a randomized block with 3 replications and triplicate

analyses of each sample; harvest dates constituted blocks.

Dry weights were determined according to conventional pro-

cedures.


Component Distribution

Celery stalks of the Florimart cultivar were divided

vertically into top (leafy), middle, and bottom and, hori-

zontally into outer and inner portions. Extracts were pre-

pared of each sample to derive further information on the

distribution of the volatile components of the vapor profile

within the plant. Three replications were used with a

completely randomized design.


Analytical Procedures

Organoleptic

A triangular test was used for all organoleptic

















Cable 1. Storaae treatments used in Phase III.

Treatment Storage Duration Subsequent
Temperature Storage

(OF) (Weeks)
0 At harvest

1 38 1

2 38 2

3 38 3

4 38 4 -
5 38 2 1 day at 700

6 38 4 1 day at 700

7 38 2 8 days at 500

8 38 4 8 days at 500

9 45 2 -

10 45 4







measurements (38). Organoleptic comparisons were made be-

tween freshly harvested celery and celery stored 2 weeks at

380F. All organoleptic tests were conducted with a con-

sumer-type panel composed of staff personnel. Panel scores

were obtained in a taste panel room with subdued lighting.

Samples were taken by dicing the center one-third of each

inner petiole, excluding all heart petioles, from 20 stalks.

Petioles with great visual differences were excluded.


Extraction

Celery petioles were washed and stripped of leaves.

All petioles were ground in a Waring blendor and the juice

expressed by hand. Each extraction sample consisted of

4,000 ml of juice which was taken from approximately 12 lb.

of petioles. The juice was introduced in 400 ml portions

into the evaporation chamber of a Nester-Faust model 500

rotary spray evaporator. The residue was removed from the

\ evaporation chamber before each new aliquot of juice was

introduced. The evaporation chamber was immersed in a water

bath at 70 C and maintained at a pressure of 30 mm of mercury,

while the condenser and collection chamber were maintained

at 0 0 C and 30 mm of mercury. Preliminary procedures differed

in that traps for collection of the flavor components were

at temperatures of dry ice-acetone and liquid nitrogen and

were placed after the cold water condenser (14, 16).

The clear aqueous condensate containing the volatile

flavor constituents was retained for solvent extraction.





13

Dichloromethane, amounting to approximately 10 per cent of

the volume of essence, was shaken with the essence of 2

minutes. The samples were washed 2 times and dried with

sodium sulfate according to conventional methods. Solvent

was partially removed in a rotary evaporator at 1040F with

a reduced pressure of 660 mm of mercury. The residue was

removed from the rotary evaporator when approximately 10 ml

remained in the flask. Further removal of the solvent was

accomplished at room temperature and atmospheric pressure.

Head-space samples were prepared by placing 500 grams

of diced celery into a 1,000 ml erlenmeyer flask. The

flasks were sealed and placed in a water bath at 55C for 2

hours. At the end of this period, the head-space gas was

taken by syringe for direct injection into the chromatograph.

Standards were prepared by placing known quantities of the

various chemicals on tissue paper and positioning in the

center of the flask's contents.


Chromatographic

All chromatograms were prepared on an Aerograph Model

600 D chromatograph equipped with a flame ionization detec-

tor. Injection sample size and instrument parameters were

maintained constant for each experiment. The sample volumes

were 0.5 ul and 2.5 ml for solvent residue and head-space

samples, respectively. The injection port was maintained

at 2350C. Helium was used as a carrier gas with an inlet

pressure of 40 psi and flow rate of 20 ml/min at room






14

temperature. A manual matrix programmed sequence was used

from 100C to 2400C at 3/min as shown in Table 2. Hydrogen

flow was 20 ml/min while the 'air was maintained at 250 ml/min.

Chart speed was in/min.

Treatment comparisons were made on 12' x 1/8"' copper

columns packed with 5 per cent W/W Apiezon L and Carbowax

20 M, both supported on 80/100 mesh Chromasorb G.

Subtractive chromatography for alcohols and aldehydes

was performed by procedures described by Ikeda, et al. (27)

and Allen (1).

Infrared analyses of certain compounds were performed

on a Perkin-Elmer Model 237 spectrophotometer equipped with

beam condenser. Spectrophotometer cells were used with .005 in

spacers and methylene chloride as a solvent.

Statistical analyses of paired samples were computed

by employing students' distribution (41). Treatment means

were compared by using Duncan's multiple range test (11)

following analysis of variance.


















Table 2. Oven temperature and time relationships for the
matrix sequence used in all chromatographic rcasurements.


Time Temperature Program power*
Cumulative int-erval

(Minutes) (C)
0-10 10 100 50

10-25 15 120 50

25-35 10 140 50

35-50 15 160 50
50-58 8 180 60

58-73 15 200 60

73-77 4 220 60
77--- -- 240 60


*Power indicator for programming rate
600 D oven.


of Aerograph Model













RESULTS AND DISCUSSION


In 1966, preliminary extractions were made from

celery stalks of cultivar Florimart according to the pro-

cedures previously discussed. Figure 1 shows chromatograms

obtained from the extracts of the liquid nitrogen and dry

ice-acetone traps. All peaks are measured at range 1, at-

tenuation 8, unless otherwise designated. Little difference

was observed between the component distribution as an

effect of the two trapping procedures (liquid nitrogen vs.

dry ice-acetone). A larger quantity of extract was obtained

from the dry ice-acetone trap than from the liquid nitrogen

trap.

High boiling components are designated as those

which have retention times greater than 50 minutes: the

time at which the oven temperature is set at 1800 C. Low

boiling fractions are those components which emerge from

the column in less than 50 minutes. Figure 2 shows a chro-

matogram prepared from an extract sample of the aqueous

condensate procedure described earlier. Reference to

Figure 1 shows that there is a proportionately smaller

quantity of high boiling components in the chromatograms

from the dry ice-acetone and liquid nitrogen traps than in

the chromatogram from the aqueous condensate shown in





























2r
-I





















: cr
* .; -- -<








* j

.4.




0
















o
i I I.'.I 17











oo

















S. -
-* *


:-_ c:
,~c ; i .












E
: : S 4 .,


v0
..)
V
*4 _,
















0
---4
'. ', f ~ ",'

.). -,as

'n a *p








~ .to


--I- : .
s ,



"j-- _^' :"

---- r-^ ^^ S **-


! I '






































0
<_'i

i -





















cz



.0

Cd



C a,
a .0


























P4
aS
a



aa
i ~ T ; R :








h0~
i b iic ^ '0








F-. a,
I ____ 01






F
Ssa K
* I, i .I; i 1 "







i -i y as

P4

a,


* .r a S
W0
r l--- i

! -~r ^; ^_ ^ ^ *:




| S- ^





a,O





'-* I4


|I i





19

Figure 2. Peaks with retention times of 72 minutes or more

made up more than 40 per cent of the total peak area of the

chromatograms prepared from these aqueous extracts, while

they made up less than $ per cent of those from the liquid

nitro,-:n and dry ice-acetone traps. While both methods

yielded many low boiling compounds, trapping with liquid

nitrogen and dry ice-acetone gave a larger proportion.

The data obtained from the chromatograms prepared

from the aqueous collections revealed several additional

peaks which had retention times greater than 90 minutes, on

the Apiezon column, when the instrument was isothermally

controlled at 2400C subsequent to the sequence listed in

Table 2. The retention times and average peak heights of

these peaks are presented in Table 3. These peaks are not

considered in treatment comparisons.

In order to establish qualitative meaning of the chro-

matcgram in Figure 2, an odor characterization was estab-

lished according to the odors noted at the exit port of the

column during the programmed sequence. These odors are

presented in Table 4 with their estimated strengths. Few

celery-like odors were observed in the low boiling range,

while a large proportion of the odors in the high boiling

range were characteristic of celery. It should be noted

that additional odors may have been present but were not

detected due to concentration or sensitivity effects.

Since many of the phthalide compounds which impart

the typical aroma of celery are sterioisomers (33), it is



















Table 3. Retention times on Apiezon L column and
average peak heights of peaks with retention time
greater than 90 minutes.


Retention time


(Minutes)
105


Average peak height


(Millimeters)
11


108

120

130

142

147

156

170

214

272

















Table 4. Odors detected at the exhaust port of an Apiezon L
column after injection of 1 ul of celery solvent extract
residue.

Time Odor* Time Odor

(Minutes) (Minutes)
11 Turpentine (f) 40 Fishy (f)
12 Woody (f) 41 Undescribed (f)
18 Orange (m) 44 Diesel exaust (m)
20 Spinach-sweet (m) 45 Celery (f)
22 Bitter weed (f) 55 Sweet (f)
24 Sour milk (f) 56 Orange peel (m)
25 Plastic glue (m) 58 Cotton (s)
26 Bananas (s) 59 Cotton (f)
27 Green bananas (m) 68 Apple (s)
28 Milk weed (f) 72 Celery (f)
29 Celery (f) 73 Celery (s)
30 Undescribed (f) 74 Apple (f)
31 Undescribed (f) 75 Celery (m)
32 Undescribed (f) 78 Cooked celery (m)
33 Undescribed (f) 79 Celery (m)
36 Undescribed (f) 80 Celery (m)
37 Mint (m) 81 Celery-carrot-like (m)
38 Terpinyl acetate (s) 82 Celery-like (s)
39 Undescribed (f) 85 Cooked celery (f)

"Odor descriptive terms selected by author.
**(f), (m), and (s) designate relative strength of odors as
faint, medium,and strong, respectively.





22

difficult to determine which compound is of most importance.

Most of these isomers readily convert to butyl phthalide, re-

sulting in butyl phthalide contamination in any isolation.

A sample of n-butyl phthalide which strongly yielded the

characteristic aroma of celery was obtained.1 This compound

had a retention time of 77 minutes on Apiezon L and 69 min-

utes, 30 seconds on Carbowax 20 M. In order to substanti-

ate further the true structure of this compound, an infrared

spectrum was prepared. The trace of the infrared spectrum

(Figure 3) very closely resembles that presented by Mitsu-

hashi, et al. (31) for Ligustilide, a sterioisomer of butyl

phthalide. However, the spectrum of Figure 3 shows absorp-

tion at 3,030 cm-1 with 1,600 and 1,535 cm-1 indicating the

presence of all three double bonds around the benzene ring

(3). These data suggest that the compound is n-butyl phtha-
lide rather than Ligustilide or a mixture of neocnidilide

and butyl phthalide (33).

On the basis of these data it is assumed that the

peak appearing at 77 minutes' retention time from solvent

front of Apiezon L is n-butyl phthalide. Collection of the

actual peak would have been desirable; however, needed facil-

ities were not available.

When using a fractional collection apparatus, Gold

and Wilson (14) found 25 compounds present in the dry ice


United States Department of Agriculture Fruit and Vegetable
Laboratory, Winter Haven, Florida.


















0

o ii






Figure 3. Infrared spectrum of compound believed to be
n-butyl phthalide.








trap. Among these compounds they discovered no acids,

aldehydes or phenols. They also indicated only 4 compounds

were present in the liquid nitrogen condensate, and the

principal odor constituent of the dry ice trap was cis-3-

hexen-l-yl pyruvate, and that of the liquid nitrogen trap

was diacetyl (16). The presence of more compounds in these

extracts seems to indicate the fractionating system of Gold

and Wilson was more efficient in reducing the vapor tempera-

ture than the system used in these experiments.

The importance of phthalides and their derivitives

as flavor contributors in celery have been pointed out by

several authors (2, 7, 14, 15, 16, 19, 31). It therefore

seems essential that a representative celery extract contain

a portion of these high boiling phthalide compounds. On the

basis of previous investigations (16, 26), the low boiling

compounds shown in the previous chromatograms were believed

to be primarily C10 H16 hydrocarbons and related compounds.

The data obtained from the aqueous essence sample are

in agreement with those of Gold and Wilson (16). By separa-

ting the high boiling and relatively low boiling fractions,

they found the phthalides primarily responsible for the

celery odor located in the column bottom. They did, however,

indicate that the addition of this material to tomato juice

did not reproduce the tomato-celery juice blends unless

material from the dry' ice or liquid nitrogen traps was

included. Since the solvent extract prepared from the

aqueous essence contains both fractions, it is believed







that the profile shown in Figure 2 most adequately repre-

sents the flavor profile of celery.


Organoleptic Evaluation

Triangular taste test comparisons between freshly

harvested celery and celery stored 2 weeks at 38 F indicated

there was a significant (0.05 level) difference between

treatments. Celery used for this first test was harvested

on April 24 (stored 2 weeks) and May 9 (fresh). However,

when comparison was made between celery harvested May 9

(stored 2 weeks) and May 24 (fresh), no significant differ-

ence was observed. In the first test 61 per cent of the

judges paired the samples correctly,while in the second test

only 39 per cent were able to pair them correctly. From

these data it is difficult to determine if differences did

exist as a result of storage or if the differences were

inherent within each harvest. Hall (21) found that the

flavor of celery did change while in storage at 40 F, and

that there was an interaction between storage temperature

and petiole position on the organoleptic measurement. When

comparing Florida-grown with California-grown celery,

variations were found between April and May harvests (22).

These differences were attributed to a change in sucrose

content and not to any large change in the volatile flavor

constituents. However, Hall (23) indicated that a taste

panel could easily be influenced by the presence of the

bitter substance which is associated with the outer petioles.





26

It is assumed that if differences do exist in the volatile

flavor constituents between celery stored 2 weeks at 38F

and freshly harvested celery, storage at higher temperatures

and for longer durations should serve to compound these

differences. Texture changes in celery stored at higher

temperatures make triangular taste tests quite difficult.


Storage

Phase I

Chromatograms used for computation of these data are

similar in appearance to the chromatogram shown in Figure 2

and were prepared using the aqueous trapping procedure.

There seemed to be little change in the chromatograms from

the celery stored 2 or 4 weeks at 380F when compared to their

respective controls.

The importance of the odor characteristics of the

high boiling phthalides has already been discussed. How-

ever, any true flavor change will be a result of the inter-

actions of all compounds present, whether in threshold or

sub-threshold concentrations (18). Changes in the ratio

of high/low boiling components will be used as an indicator

of change for all chromatograms. Changes in this ratio do

not necessarily indicate flavor changes since no research

has established a correlation.

The average ratio of high boiling/low boiling com-

ponents for all chromatographed samples was 1.32. Ratios

for high boiling/low boiling fractions for the low temperature





27

storage treatments are presented in Table 5. ".e ratio for

the freshly harvested sample (1.33) was very near the aver-

age for all treatments. While no significant chrsnge was

observed between freshly harvested samples and those stored

2 weeks at 380F, there was a slight increase in the mean

ratio after 4 weeks' storage.


Table 5. The effect of low temperature (38 F) storage on
the ratio of high/low boilinr fractions in celery flavor
extracts.


Treatment
Harvest 2 weeks'38F 4 weekst38F

(Fraction ratio-high/low)
1.33 1.30 1.65


While a large number of peaks were resolved for each

chromatogram, only a few contribute substantially to the total

peak area of these chromatograms. The 9 peaks which contrib-

ute most (more than 1.0 per cent of total area) to the total

peak area are listed in Table 6. Also presented in this

table is the mean per cent each contributed to the chr-oato-

gram total for the respective treatments. The only signif-

icant change observed was an increase in the peak at 17

minutes (tentatively identified as d-limonene) when celery

was stored 2 weeks at 38 F. This is difficult to explain

since the mean for this treatment is higher than the control

and the sample taken at 4 weeks' storage. It is noted that

there was a large increase in the mean per cent of total area














Table 6. Th- effect of storage at 38 F for 2 or 4 weeks
upon the relative per cent of 9 major chromatogram
components.

Retention Treatment
Time At harvest 2 weeks 4 weeks

(Minutes) (Per cent)
17 20.3a* 23.4b 20.2a
24 6.9 5.6 6.7

36 3.5 2.5 1.7

56 1.0 1.1 1.0
62 1.8 2.7 2.7

72 11.3 7.7 8.4

78 23.6 31.5 29.9
80 9.9 4.4 7.5
81 3.1 3.4 3.6

(Total per cent)
80.4 82.3 81.7


*Values with differing letters in horizontal rows are
significantly different at the 0.05 level according
to Duncan's multiple range test.







for the peak at 78 minutes when the celery was stored 2 or

4 weeks as compared with the control. This increase was

reflected in an increase in the ratio of high/low boiling

components for the 4 weeks'storage treatment; however,

increases in the low boiling fraction negated this effect

at 2 weeks'storage. No significant difference was observed

between the total per cent represented by these 9 peaks.

Chromatograms were prepared from celery extracts at

weekly intervals during storage for 10 weeks at 380F. After

celery had been stored 8 weeks at 380F it was of no market

value, and approximately 25 per cent of the petioles had to

be removed before the extracts were prepared.

Figure 4 shows a chromatogram prepared from celery

stored 10 weeks at 38F versus freshly harvested celery.

Peaks in areas of highest concentration (12 to 22 minutes

and 72 to 90 minutes) seemed to have been maintained while

peaks in areas of low concentration (24 to 70 minutes) have

been reduced or are missing from the chromatogram. Note-

worthy is the fact that the low boiling components are pre-

sent in abundance after storage for long periods of time and

regardless of the condition of the celery.

Observation of chromatograms prepared from freshly

harvested celery show no peaks of large area between 85 and

95 minutes, retention time. Also, no build-up in concentra-

tion was experienced in this range when celery was stored 2

or 4 weeks at 380F. However, when celery extracts from

celery stalks stored 8 weeks or more at 380F were





30






i | c4


S .. .. *


. -. I '* W












- ri4


1
.. 4)P ; ; "" ?


<* .S 02




V :
K >1
Vl










0--





31

chromatographed, 2 lri-io (peak height greater than 300 mm)

peaks were present at 87 and 91 minutes' retention time.

Celery stored 5 days at 70F was quite dehydrated

and of poor market quality. The freshly harvested control

for this study had a high/low component ratio of 0.98, which

was lower than the average. This ratio increased to 2.45

when the celery was stored 5 days at 700F (Figure 5).

This increase in component ratio was accompanied by

a corresponding highly significant decrease in total peak

area when compared to the freshly harvested control. The

mean peak area for the control chromatograms was 48,162.3

mm2 while the mean peak area from the chromatograms of the

70 F storage treatment was only 23,296.5 mm2 (Figures 6and

7). Such a large change in peak area would not normally be

expected because the extracts were prepared from a constant

volume of juice. Since the moisture content of the celery

decreased during storage, it is possible that under these

dehydrated conditions a larger proportion of the volatile

components remained in the pulp and were not extracted.

However, Matthews (30) found a reduction in total volatiles

of beans after storage for 5 days at 70 P. A conversion of

the volatile flavor components to,non-volatile forms through

enzymatic action would result in such a change. However,

no data have been collected to substantiate this hypothesis.

The changes that occurred in the extracts prepared

frcm celery stored 5 days at 70 F can be seen in Figure 7.

There was a decrease in the peak area of all peaks,
























3.0

0


~ 2.0
o
.-i


1.0






Control 5 days at 700F


Figure 5. Mean ratio of high/low boilingocomponents
as affected by storage for 5 days at 70 F (differ-
ence significant at 0.05 level).
























D 40-
o
0o
0O


. 30



o 20.



10




Control 5 days at 70F


Figure 6. Mean peak area of chromatograms prepared
from celery extracts taken at harvest and after 5
days' storage at 700F (difference significant at
0.01 level).












---- --------~-- -------


I ~
I



1;, P


.=. .. .. .. .. .. ---



Ii-- I -T






^ ^- .. .... .... 4








- T
4.

1--
























S *
I 2 -Ai -











' .- ... s'






* *-- .. 0- A



--













S Cd
-- = Sa--







'-----^*r **
i -l0

--Os-
.~ V
ac





35

particularly those in the low boiling range which appeared

before d-limonone (17 minutes). The high boiling components

were moderately stable and the peak area decreased only

slightly.

Distribution data of the 9 large peaks from the chro-

matograms prepared at harvest and after 5 days' storage at

70F are presented in Table 7. There was a significant

decrease in the proportion of the total peak area repre-

sented by the peak at 17 minutes. Noteworthy is the fact

that this change is opposite that observed when celery was

stored at 380F. However, there was a corresponding increase

in the per cent represented by the large peak at 78 minutes

when the celery was stored 5 days at 70 F. From these data

it seems likely that the components of high vapor pressure

were metabolized or lost through evaporation as the celery

was dehydrated. Table 7 also presents the per cent of the

total peak area represented by the sum of the peak areas of

the 9 major peaks. The sum of the areas increased when

celery was stored 5 days at 700F. Reference to Table 6

shows that there was a similar increase when celery was

stored 2 and 4 weeks at 38F. This would suggest the pos-

sible elimination or decrease of many of the small peaks

not considered in these data.


Phase II

Celery used for the market simulation study had a

very high peak ratio of high/low boiling components at















Table 7. The effect of storage at 70 F for 5 days upon the
relative per cent of 9 major chromatogram components.

Retention Treatment
Time At harvest 70F for 5 days

(Minutes) (Per cent)
17 19.5 9.0'
24 5.7 7.5

36 3.3 1.4
56 1.0 1.6
62 2.1 4.1

72 13.6 8.0
78 22.0 36.9
80 9.3 7.2
81 3.3 9.9

(Total per cent)
79.8 85.6

*Difference significant at the 0.05 level.





37

harvest as compared to the average for all treatments. The

ratio was 2.38 for the freshly harvested celery and decreased

significantly (0.05 level) during storage to 1.09 (Figure

8). The peak areas for the treatment moans used in com-

puting the high/low ratio are shown in Figure 9. The dif-

ference between the areas of each treatment are 21, 39, and

19 per cent for low, high, and total fraction areas, respec-

tively. While there was change in both the low and high

component areas, these data indicate a greater change in the

high fraction.

A comparison between the mean peak areas of two com-

pounds, d-limonene and n-butyl phthalide (identification

based on retention times of known compounds on polar and

non-polar chromatographic columns), extracted from the con-

trol and market simulated lots of celery,is presented in

Figure 10. There was a highly significant increase in the

peak area of d-limonene in the chromatograms of the market

simulation treatment when compared to those of the control.

Considerable variation was observed in the areas of the peak

representing butyl phthalide; however, there was a substan-

tial decrease in the mean peak area after storage under

market simulated conditions. The chang.js observed for these

two components were in opposite directions.

Any comparison between limonene and butyl phthalide

is quite critical from an organoleptic point of view. Th3

importance of the high boiling phthalides has been stressed

by the odors noted in Table 4 and by the research of other




















2.4


2.0


0




0
o
p 1.0










Control Market simulation


Figure 8. Mean ratio of high/low boiling components
of the chromatograms from the control and market
simulation treatment (difference significant at
0.05 level).




















80


70 -


Control



SMarket simulation


Low High Total


Figure 9. .*'ean peak area of low and high boiling frac-
tions of the chromatograms of the control and market
simulation treatment.






40











d-limonene


30 n-butyl phthalide


25- N
0
0
20-


A15


< 10_







Control Market simulation


Figure 10. Mean peak area of d-limonene and n-butyl
phthalide for the control and market simulation
treatment (k*-:change in peak area significant at
0.01 level).






41

investigators (14, 16). Terpenes, however, seem to be less

dominant in determining vegetable flavor than in many of the

fruit flavors. It is difficult to determine whether a

decrease in the high boiling fraction during storage would

result in a decrease in the actual potency of the celery

flavor. Any discussion of the decrease in proportion of

these elements should take into consideration the additive

effects of the various sub-threshold concentrations which

are affecting the flavor profile (18). This additive effect

of the various components further complicates the discussion

of the actual organoleptic effect of a proportionate change

in high or low boiling components.

Grouping other peaks of the chromatograms seems to

substantiate the change in limonene in the low boiling range.

When peaks in the low boiling range (excluding limonene) were

considered as a whole, little change was observed between the

control and the market simulated study (Figure 11). There

was a considerable decrease in the peak area total for those

peaks which occurred in the high boiling range (excluding

butyl phthalide) as a result of storage under market condi-

tions. It is quite possible that an imbalance of these com-

ponents at harvest stimulated a change to a more balanced

situation. It has not been determined if there could be a

conversion of high boiling components to low boiling com-

ponents during storage, but enzymatic processes are active

in the formation of volatile flavor components (25).

In addition to a slight decrease in total peak area


















SLow boiling components


30 N High boiling components


0

S\0



15
o \

< 10







Control Market simulation


Figure 11. Mean peak area of all peaks in the
low and high boiling range excluding limonene
and butyl phthalide, respectively.






43
as a result of storage undor market conditions there was a

hir'-.ly significant decrease in the proportionate amount of

this total represented by the 9 previously mentioned peaks

('_ble 8). Also shown is a decrease in the proportionate

amount contributed by each of the 9 peaks except peaks at

17 minutes (d-limonene) and at 81 minutes. However, indi-

vidual peak changes of substantial importance are those at

17, 72, and 78 minutes. Others in combination mifht also

contribute substantially as noted above. "Thse data also

indicate a net decrease in those high boiling flavor com-

ponents most -rsponsible for the aroma of celery.

Chromatograms prepared from the celery extracts of

the control and the market simulation treatments are con-

tained in Figure 12. As was pointed out previously, there

was an increase in the low boiling fraction, particularly

limonene, and a proportionate decrease in the high boiling

fraction between 72 and 86 minutes. When celery was stored

at 45 F for 2 weeks and 50F for an additional week, 2

large peaks appeared in the chromatograms at 87 and 91

minutes. The 2 peaks shown in the market simulated treatment

chromatogram at 87 and 91 minutes have peak areas of 1,305
2 2 2 2
mM and 2,004 mm as opposed to 72.3 mm2 and 176 mm for the

chromatogram prepared at harvest. These data are in agree-

ment with those received from chromatograms of celery ex-

tracts prepared from celery stored 8 to 10 weeks at 380F

(Figure 4).

Considerable variation occurred in peaks on these














Table 8. The effect of storage under market simulated condi-
tions upon the relative per cent of 9 major chromatogram
components.


Retention Control Market Change
Time Simulation

(Minutes) (Per cent)
Low boiling
17 16.6 28.7 +72.9*

24 3.6 3.6

36 2.8 2.4 7.1
High boiling
56 3.4 2.8 -17.6
62 5.8 4.8 -17.2

72 11.9 9.6 -19.3

78 33.5 7.7 -77.0

80 9.8 8.9 9.1

81 1.5 2.0 +33.3

(Total per cent)
94-4 80.1**


*Significant at 0.05 level
**Significant at 0.01 level














~__~_-/
i---~_


c-_~
7;;"


i
R-e
P~
;Ii



cc










3


..
sr^--__












a ^



















.---- --.--- --




o
~ \ 2







0


-2








;4
0
o



ct
r3


o
02




fc

o<
0








o o
S0






t-0







0 '




?< i


S8







chromatojrams between 37 and 45 minutes'retention time.

Pea'rs in this area were always present in measurable con-

centrations at range 1, attenuation 8; however, the relative

peak areas were not constant within replications.

while e the celery used in Phase I and Phase II was of

the same chronological age, it is doubtful that the physio-

logical maturity was the same. An indication of this vari-

ation could be observed differences between the ratio of

high/low boiling components for the freshly harvested
0
samples used for the 70 F storage treatment and for the

market simulation. Celery used for the market simulation

was harvested in July while that used for the 70 F treatment

was harvested in May. If the physiological maturity was

not the same for all treatments, generalized comparisons

cannot be rendered between treatments (9, 12, 13, 24).


Phase III

Nine separate peaks were resolved from Florida 2-13

celery using the head-space analysis technique. Identifi-

cation of 4 of these compounds was accomplished by compari-

sons of their retention times on polar (Carbowax 20 M) and

non-polar (Apiezon L) columns and by spiking these chemicals

into flasks containing the celery for head-space measurement.

Since all comparisons are made with chromatograms from the

Apiezon L columns, numbers have been assigned to the re-

solved compounds on the basis of retention time on this

phase (Figure 13). The chromatographic profile obtained






47





















6







5

2 7





0 5 10 15 20 25
TIME(min.)





Figure 13. Chromato:7ram from head-space measurements of
celery volatiles with peak numbers according to reten-
tion time (Apiezon L).





48

-hen these samples were chromatographed on Carbowax 20 M is

shown in Figure 14. Comparison data were prepared on the

basis of an instrument sensitivity of range 0.1 and attenu-

ation 16, while peaks 6 and 8 were measured at attenuation

32 and 64, respectively, for both columns.

Observed retention times of the 9 volatile constit-

uents are presented in Table 9. These are measured from

injection and are somewhat different from those expected

when pure compounds are chromatographed. This deviation in

time is an effect of the large amount of water vapor present

in the holding flask (34). However, the relative retention

times are not greatly affected by the presence of water.

The retention data presented in Table 9 are in agreement

with those of Hunter and Brogden (34). These investigators

found cA-terpinene to be present in volatile extracts and

the location of this compound in their chromatograms sub-

stantiates the idea that it is responsible for peak 9.

However, no retention data have been collected to support

this hypothesis.

The relative per cent of total peak area of the 9

measured peaks is presented in Table 10. Approximately 88

per cent of the total peak area of these chromatograms is

accounted for by the four Co1H16 hydrocarbons identified,

while d-limonene and ,/-pinene resulted in approximately 83

per cent of the total peak area.

The mean peak areas (fresh weight basis) of the 9

peaks are listed in Table 11 according to treatment. Total



































4






0 5 10 15
Tl ME(min)




Figure 14. Chromatogram obtained when celery head-space
volatiles were chromatographed on Carbowax 20 M.

















o 0 I I 0 (M I I-
I .. .. .. I ..
Cj '. __ s r r-
0 0

c)






o H 0 0


0
P4 r CU o
Or 0 0 r- c(O 0CJ (M r(k (










E 0 t .. .. .. m ..




HL2 C





0 0 Go tf\ 0 0 'VN 0 0
0 0 0t _-r UN 0 I\ CJ
0, -a 0 o




4a4









0
Q00 1
o c























H (Z CD r co 0o
0 I 0 I






















:, "O U d 'Q '2
!D rc; ;









eI K) P-
E 2



















Table 10. Relative per cent of total peak area for
the 9 peaks measured in head-space analyses.


Peak number Per cent of total area

1 0.30

2 2.30

3 0.18

4 1.46

5 2.84
6 11.43

7 3.90
8 72.23

9 5.30







Table 11. Mean peak area (fresh weight basis)for each of
the 9 peaks analyzed in head-space measurements as related
to storage temperature and duration.


Storage
Treatment Peak number
1 2 3 4

(Area mm2)


Control


16. 5a


277.4a


6.3a


122.2a


1 wk @ 38 F


2 wk @ 380F

+**1 day @ 70F

+ 8 days @ 50F


2 wk @ 450F


19.5b


43.3e
37.6c

52.4g


47.6f


310.4ab 12.3a


407.5ab
296.5c


20.4bc

16. labc


310.Oab 25.3c


441.lb


27.8c


145.9ab


263.5cd

177.3ab
227.1bcd


251.7cd


3 wk @ 380F


4 wk @ 38F

+ 1 day @ 7 0
+ 8 days @ 50F



4 wk @ 45F


44.6e


389.8ab 25.8c


40.2d 258.7a 22.6bc

39.3cd 309.lab 24.3bc


58.3h


411.0Oab


389.lab


21.8bc


25.7c


228.0bcd


199.0abcd

181.4abc


281.3d


269.8d


*Those means in vertical columns not
letter are significantly different


followed by the same
at the 0.05 level.


+ designates subsequent storage treatment after
at 380F.


storage







Table 11. Continued


Stora.ge
Treatment Peak number
5 6 7


Control


1 wk @ 38 F


2 wk @ 38 F

+*'1 day @ 70F

+ 8 days @ 50 PF


2 wk @ 450F



3 wk @ 380F


4 wk @ 38F

+ 1 day @ 70 F

+ 8 days @ 50F


4 wk @ 45F


306.Oab


497. c

412.5bc

464.9bc


521.1lc



500oo.5c


406.7abc

388.9abc

494.5c


500.8c


(Area mm 2)

1,187.7a



1,430.2ab


1,814.5c

1,532.6abc

1,803.0bc


2,141.2c



1,930.0c


1,514.2ab
1,649.0abc

2,137.5c


1,973.8c


Those means in vertical columns not followed by
letter are significantly different at the 0.05
*'+ designates subsequent storage treatment after
at.380F.


the same
level.
storage


304.9ab


461.2ab


875.9b

593.4ab

642.5ab


732.lab



660.8ab


255.8a

369.0ab

766.3ab


708.3ab







Table 11. Continued


Storage Peak number
Treatment 8 9 Total(l-9)

(Area mm 2)


Control


1 wk @ 38 F


2 wk @38 F

+**1 day @ 70F

+ 8 days @ 50F


2 wk @ 45F


3 wk @ 38F


4 wk @ 380F
+ 1 day @ 70 F

+ 8 days @ 50 F


4 wk @ 450F


7,002.3a


l1,650.8a

11,362.2a

12,149.8a


14,232.3a


11,440.3a


11,986.7a

10,501.9a

12,118.4a


14,064.7a


552.9a


571.6abc


897.1abcd

688.0abc

804.1abcd


1,069.2d


826.6abcd


760.7abcd

666.1abc

1,459.2e


899.lcd


8,144.la


10,260.8ab


16,456.4ab

15,117.lab
16,481.5ab


19,475.4b


16,030.lab


15,389.5ab

14,150.2ab

17,348.0Oab


19,381.4b


*Those means in vertical columns not followed by the same
letter are significantly different at the 0.05 level.
+ designates subsequent storage treatment after storage
at 38 F.








area for these peaks as a result of treatment is also pre-

sented in this table. While all values were not signifi-

cantly different, the mean peak area for the freshly har-

vested control was lower than any other treatment for all

peaks except peaks 2 and 7. The lowest peak area for these

peaks occurred when celery was stored 4 weeks at 38 F.

Statistical differences were observed between treatments for

all peaks except peak number 8 (limonene).

Without exception, the mean peak area of all peaks

progressively increased as a result of storage at 380F for

1 and 2 weeks as compared with the control (Figure 15).

This increase in peak area continued for some compounds

until the third week of storage; however, all peaks except

peak 8 showed a decrease in peak area after 3 weeks' storage

at 380F.

Figure 16 shows a comparison of the volatile changes

(FWB) resulting from storage of celery for 1 day at 70 F

and 8 days at 50 F after storage for 2 and 4 weeks at 38 F.

Considering only the treatments subsequent to 2 weeks' storage

at 380F, without exception, all compounds from the 70 F

treatment had a lower peak area than those from the 50 F

treatment. There was only one exception to this trend (peak

3) when 70 F and 50 F treatments were prepared subsequent to

4 weeks'storage at 380F. Also, storage of celery for 1 day

at 70 F after 2 weeks at 38 F resulted in a decrease in peak

area of all peaks as compared to the freshly harvested

sample. The results of the same treatment after 4 weeks at





























Figure 15. Trends in the change of peak area (fresh weight
basis) as a result of duration in storage at 380F.












PEAK


x


I--I~CI


~---------------


------------------ --- 1


-
II --C--


A'---~


WEEKS


---------9


r --"---cl

















0202

a2) Vd

9 md
co




cX2





q4GO
ca
P-4
('Jo
Q0
C d
CUO

O

4a)


04

<00
4->






CO
C.)








0n b-
0rl
ca ^

















,CO
v0
m o

















-02a



0 qto0
r1 OS'0

o 0
U
h f













lza
0N4
02i-I

(0
02 tJo



.0 (N)
+1 -P







O


:




1: 0 0
-3 -
o cG cd

(D mp,









Xi-
C O CC.
cbOS
ca 'a




P-, -
-Pos c




01
'.0 'a~

aCO -




P40
+.t-I

















5 ww






(0otx)
d-~---- -4





(OOLX)










r- C'. -


C w
(OOkx)







4 h
N O






































to It) 3
IF 5 f 5 i


c*BC cy;F*


cu rt^


5 I^ i
F ~ ^ r




















Sio ~ in a '' il
(ZWYU) V3ayv


(004 Y)









S3;





(000 X)



L 1 U. ib

V>,
S -


+ +


(0 r- 1 n1
(O0 x)





e t.1--3








(00lX)
(Zww)v3 v


__







380F are more variable, peaks 4, 5, 8, and 9 showing the
decrease after 1 day at 700F. In some treatments, storage
for 8 days at 50F resulted in an increase in peak area
when compared to the corresponding 380F treatment. These
effects were, however, more prevalent after 4 weeks at 38 F

(all peaks except peak 2 resulted in the increase).
Peak areas for the control and storage treatments of
2 and 4 weeks at 38 F are presented in Figure 17. There

was no significant difference between the two storage treat-
ments at 45 F. A comparison of the 45 F treatments at 2

and 4 weeks reveals that the peak areas resulting from the

45F treatment tended to be higher than those at 380F. For
both 2 and 4 weeks'storage, the only exceptions were peaks

4, 6, and 7 at 2 weeks, no exceptions being noted at 4 weeks,
storage. Noteworthy is the fact that total peak area of all
9 compounds for the 45 F treatments, both 2 and 4 weeks,
were the only total values significantly different from the
control.

The general trends for these treatments are summa-
rized in Figure 18. The points in this graph are based on
the total peak area of all peaks. Storage at 380F for 1
and 2 weeks resulted in a linear increase in peak area,
while further storage at 380F caused a reduction in peak
area. Storage at 70 F for 1 day after storage at 38 F
resulted in a decrease in peak area, while storage at 50 F
for 8 days after 38 F holding resulted in an increase in
peak area. The degree of increase at 50 0F was greatly













(D
tO
C3 qt-4
F 0
0


4p -4
m o

(D

;4 ^





coF-
I03

CI22



d02
(n
cP ^






(1)
r-I T-.
P40


co
04
02


0205
P-V





o 00
QdQ


05
hOh













+0 C

-P 02
4-') i -i








a o



s Cd

bD P


cz 1
0 (D

mMC
0 0

(D S


S:~4 02zi
~v


co 2
lz CO












I' a
Cd (\I r=

0201
.D 02



0
Cd~c
-A 00
* 0 t2

cri o ON





Q c









4-jr p4
0O CO4
X> 02
h0 P


















rX4o
? ca '
rCO
,fl *
vi f
















C,.=






(oolx)


4.




(Oix)



E' 1^r'Mr..
I'
Y, C~I(~~Ein~r^^^^^.


0 0 0






a.(










0.








6 (W 6 Va 6
(z4v4) V38V


BS0o


IA 0 IA
( o x)
(O0lX)


*cpp 1 "1



Z" is,;
013V9


(00LX)





3 S










0 0





(OO


(OOLX)












20-


15.
0
0
0




- 10-
5Q





5-


--.- 45F
-----38F
70F
......... 500F


..................
/r -
/ A
/ /
/
I I
/ /
! !


SI

~/


0 1 2 3 4 5
Duration of storage (weeks)

Figure 18. General trends in total peak area (fresh weight
basis) changes according to temperature and duration of
storage at 450F, 380F, and subsequent storage at 50OF
and 700F after storage at 380F.








affected by the increase and large value of peak area for

peak 8. The greatest total peak areas were observed from

celery stored 2 or 4 weeks at 45 F. Interpolation of the

change between harvest and 2 weeks' storage at 45 F is not

\\ possible for these data.

After one experimental block had been analyzed, there

appeared to be differences associated with apparent moisture

loss. In attempts to define these effects, dry weight

measurements were prepared from celery in the second and

third block and these data are presented in Table 12. 'Few

significant differences were observed as a result of storage

treatment. However, the per cent dry weight of celery

stored 2 and 4 weeks at 450F was significantly higher than

most other storage treatments. Also, the control was signif-

icantly different from only 2 treatments: 2 weeks' storage

at 45 F and 3 weeks' storage at 38 F.

Since these differences in dry weight were observed,

the peak areas were calculated on a dry weight basis (DWB)

and are presented in Table 13. When calculated on the basis

of dry weight, there seemed to be more variation in the peak

areas than when calculated on a fresh weight basis. The

peak area (DWB) for the control was not lowest; instead,

there was a general decrease in 4 of the compounds to 1 week

storage at 38 F (Figure 19). After this storage duration,

the peak areas of the various peaks increased as when the

data were calculated on a fresh weight basis. All peaks

increased in peak area from 1 week to 2 weeks' storage at







Table 12. Per cent dry weight of celery as
related to storage temperature and duration.


Storage
Treatment


Control


1 wk @ 380F


2 wk @ 380F


+*"*Ik day @ 700F

+ 8 days @ 50F


2 wk @ 450F


3 wk @ 380F


4 wk @ 38F
o
+ 1 day @ 70 F

+ 8 days @ 50F


Dry weight

(Per cent)

4.llbcd"


3.99abc


3.95abc

4.19bcd

3.94abc


4.82e


3.65a


3.78abc

3.67ab

4.17bcd


4 wk @ 45'F


4.60de


*Those means in vertical columns not followed
by the same letter are significantly differ-
ent at the 0.05 level.
+ designates subsequent storage treatment after
storage at 380F.








Table 13. Mean peak area (dry weight basis) for each of
the 9 peaks analyzed in head-space measurements as re-
lated to storage temperature and duration.


Storage
Treatment Peak number
1 2 3 4

(Area mm2)


Control


1 wk @ 38F



2 wk @ 38F

+**l day @ 70 F

+ 8 days @ 50F



2 wk @ 45F



3 wk @ 38F



4 wk @ 38F

+ 1 day @ 70F

+ 8 days @ 50 F



4 wk @ 450F


0.4a* 17.7ab


O.5a



1.2ab

0.9ab

1.3ab



l.Oab



l.Oab



0.9ab

1.3ab

1.6b



0.8ab


17.lab



19.0ab

16.0ab

21. ab



21.2ab



24. lb



14. 5a

18.9ab

22.4ab



18.5ab


0.3a



0.4a


0.8bc

0.7ab

1.2c



1.2c



0. 8bc



0.7ab

l.Obc


6.0abc



4.9a


9.2d

5.6ab

8.7cd



8.0bcd



8.6cd



8.0bcd

8.4cd


1.Obc 10.6d


l.Obc


8.4cd


Those means in vertical columns not followed by
letter are significantly different at the 0.05


the same
level.


+ designates subsequent storage treatment after storage
at 380F.


--








Table 13. Continued


Storage Peak number
Treatment
5 6 7

(Area mm 2)


Control



1 wk @ 380


2 wk @ 38F

+**l day S 70 F

+ 8 days @ 50F



2 wk 45F



3 wk @ 380?



4 wk @ 380F

+ 1 day @ 70OF

+ 8 days @ 50F



4 wk @ 450F


9.6ab*


8.3a



14.9cd


10.9abc

13.6bcd



15.0Ocd



15.7d



ll.7abcd

13.8bcd

15.-lcd



13.9cd


68.9ab



60.6a



72.2ab

64.2ab

74.8ab


82.0ab



92. 5b



67.lab

72.7ab

91.6b



75.2ab


9. 8ab



5.3a


10.8ab

8. 2ab

9.9ab


12.Ob



11.3ab



8.2ab

9.7ab

13.3b



11.6ab


*Those means in vertical columns not followed by the same
letter are significantly different at the 0.05 level.
+ designates subsequent storage treatment after storage
at 380F.







Table 13. Continued


Storage Peak number
Treatment 8 9 Total(l-9)


Control


1 wk @ 38F


2 wk @ 38F
+**l day @

+ 8 days @


2 wk @ 450F


3 wk @ 38F


4 wk @ 380F

+1 day @ 70F

+ 8 days @ 50F


4 wk @ 450F


700F

50F


(Area

233. 2ab*


164.3a


232.4ab

251.7ab

279.8ab


318.5b


301.6b


293.7b

286.2b

315.4b


321.4b


mm2)

30.4abc


23.4a


37.5cd
27.8ab

34.1bc


45.2d


36.0bcd


31.labc

37.0bcd

35.6bcd


37.6cd


376.4ab


284.8a


397.9ab

386.0ab

444.5b


505.2b


491.6b


431.6ab

450.6b

506.7b


488.6b


Those means in vertical columns not followed by the same
letter are significantly different at the 0.05 level.
+ designates subsequent storage treatment after storage
at 38 F.































Figure 19. Trends in the change of peak area (dry weight
basis) as a result of duration in storage at 38 F.






70



300- PEAK
"-- NO. 8

250"


200-


150
95


85


75


-j 65 -6











25
15 255

45
K *9



0.25 --








158 ---.-----a--- 3





0.2


0 1 2 3 4
WEEKS





71
380F. Also, all peaks showed a decrease in peak area from

3 to 4 weeks when storage was at 38F.
Figure 20 shows a comparison of the volatile changes
(DWB) resulting from storage of celery for 1 day at 70F

and 8 days at 50F after storage for 2 or 4 weeks at 380F.
Considering only the treatments subsequent to 2 weeks' storage

at 38 F, without exception, all peaks from the 70 F treat-
ment had a lower peak area than those from the 50 F treat-

ment. This trend was not true for peaks 3 and 9 when celery

was sampled after 4 weeks at 380F. Also, storage of celery

for 1 day at 70F after 2 weeks at 380F resulted in a de-

crease in peak area for all compounds. The results of the
same treatment subsequent to 4 weeks at 38F do not agree

with those after 2 weeks at 380F when compared with the 38 F

check at 4 weeks' storage. All peaks except peaks 3, 8, and

9 showed a progressive increase in peak area for 70 F and

50F storage, respectively. This trend was noted in peaks
2, 6, and 7 for the data calculated on a fresh weight basis.

Peak areas (DWB) for the control and storage treat-

ments of 2 and 4 weeks at 380F and 45OF are presented in

Figure 21. All peaks except 1, 4, and 5 showed an increase
in peak area when stored at 45F (2 and 4 weeks) as compared

to 380F storage. The differences between these treatments
and the control are not as great as when the data were

calculated on a fresh weight basis.
When considering the peaks on a dry weight basis, the

celery stored 2 and 4 weeks at 450F did not result in the













02
4-')
0 -

4-

0 W
q 0

Fcd
col
CO~l







0+
c'
SV\







4-,
'O CO







C 0
02 o-






0
CO



(Q



P4 C
E
co







944






0

0
o p









ca
cs












-P
COPa











blil
0O

0





4-p
(D



02




0 C











02)
rlC











o co
CDO


hm



*r0 0








Y 0
~ad













co0 PH
0V 1
0202




tJo o
+H


r,
? o
-' C


CUO !3
(D 0'-
Q) as\ m



C i-
CO C










































S-
n I



























o .







F-i ?... : ---,.- (
















~Z V3VN
0





E^.44 V
L---I .
^ ifl"

P Ws n
c----- 3


in o mO
*- c-


(, V ,


In


I I/I






i. *
w





o









ui/

B^^-.


(,k4tN) V38Y


AL I LL
LL El b. ,


.0 e 0 -

>0 0I-


L L lI















bo o
(C

020







-P
(a Q





F4





4-4
lCC





Wo
>-d-











a c
r-p
















o 0
0 0


02





co
















(a
ca(D

N 0
PC '0 c







P-i4















I i- -.




J
-.-





0to
o 10 3

















o A 0 B .
: rr--
































I n
NU7-
0- -q qd


i 0 2


Vtw 3?JY


oo
Li w




o I 0








-(j









































10
a 1, 4
























































0 0 "0 0 0
ia .ll "l
(24~ i


41 0
CIA cY








highest area for all peaks; however, the total peak area
for these treatments in each case was higher than the corre-
sponding treatment at 380F.

A summary of the net effects of these variations is
presented in Figure 22. This graph varies from that pre-
sented in Figure 18 in two ways. First, there was not a
progressive increase in the total peak area from harvest

through 2 weeks' storage at 380F when calculated on a dry
weight basis. There was, however, an increase in the total

peak area from 1 through 3 weeks' storage at 38F with a
decrease after storage for 4 weeks. Second, when calculated

on a dry weight basis, the trend of lowering peak area with

storage of celery at 700F was not consistent. The same

decrease was noted in these data as in those calculated on

a fresh weight basis when celery was stored 1 day at 70F
after 2 weeks'storage at 380F; however, when subsequent

storage at 70 F was performed after storage for 4 weeks at

380F, the peak area (DWB) increased slightly. This reflects
the inconsistency of peak 8 (Figure 20) which was the only

peak showing a decrease after 2 weeks' storage at 38 0F and
1 day at 70PF. Those treatment effects experienced at 45 F

and at 50 F for fresh weight determinations were also true
when those data were calculated on a dry weight basis.
Peak areas for 45F treatments remained higher than any
38 F treatment and treatment at 50 F for 8 days after 38 F
storage resulted in increases in peak area.
Considering the peak area on a fresh weight basis,








-- --38F
--- 45F
700 F
................. 50F


/ / .
/ /
/ / .+
/ / ..y"


4 /
/ /
o
o / /
K+
c \ /
S\
w /

o\ /







2-








0 1 2 3 4 5
Duration of storage (weeks)

Figure 22. General trends in total peak area (dry weight
basis according to temperature and duration of storage
at 45 F, 380F, and subsequent storage at 50F and 70 F
after storage at 380F.








there was an increase in volatile content from harvest to

2 weeks' storage whereas there was little change through 4

weeks' storage. In fact, all storage treatments except

those at 70 F resulted in increases in measured volatile

content at some time during storage. From these data it

appears that there is a synthesis of the CIoH16 hydrocarbons

in storage at temperatures of 38 F, 450F and 500F. However,

a loss of these hydrocarbons during storage at 70F indicates

they are metabolized or otherwise partially lost through a

physical system. It is not possible to determine from these

data what limiting factors caused a decrease in the buildup

of these components after 2 weeks' storage of celery at 38OF

and 45F.

The reactions taking place in the holding flask have

not been studied in these experiments; however, the temper-

ature of the holding flask (55C) is near the destructive

point of most enzymes. If the enzymes controlling the pro-

duction of CIoH16 hydrocarbons were functional at this

temperature, the amount of volatile components present after

2 hours' holding time would relate directly to the amount

of substrate present in the tissue.

The biogenesis of terpenes is not fully understood;

however, it does involve the condensation of isoprene units

through mevalonic acid and squalene; mevalonic acid being

formed from acetic acid in plant tissue (5). The mechanism

involves condensation of acetyl CoA and acetoacetyl CoA to

beta-hydroxy-beta-umethyglutaryl CoA. The latter then is







reduced to form the mevalonic acid. Since the precursors

of the terpenes are common metabolic agents, the synthesis

or degradation of these products should be temperature

dependent and subject to the same limitations of other

metabolic activities. A buildup of the terpenes could only

be explained as a result of the degradation of other meta-

bolites (such as beta oxidation of fats) to form available

acetyl CoA and acetoacetyl CoA. The amounts of available

acetyl CoA and acetoacetyl CoA required to influence volatile

changes are not known.

It is a known phenomenon that cutting of tissue re-

sults in an increase in metabolic activity. The injury to

the tissue in these experiments was maintained constant and

should not have affected the results.

The changes in moisture content did not greatly effect

measured peak area in the head-space experiment. There was

no significant correlation between celery dry weight and

measured volatile content as based on correlation coeffi-

cients. The increase in dry weight of celery stored 2 or

4 weeks at 450F was a result of inadequate control of

humidity, even though celery was placed in plastic bags to

minimize water loss. A larger buildup of the volatile

constituents with these treatments could have been indi-

rectly affected by the reduced moisture content.

The 9 volatile components measured in the head-space

analyses correspond roughly to the chromatographic fraction

of the solvent extracts with retention times of 8 to 20







minutes. There was a considerable decrease (Figure 7) in

the quantity of these components being measured when celery

was stored at 70OF for 5 days. Also, storage of celery for

2 weeks at 450F and 1 additional week at 50F resulted in

an increase in these components and a highly significant

increase in the amount of limonene (Figure 10). The changes

in the Co1H16 hydrocarbons in the head-space analyses

corresponded to those observed when celery solvent extracts

were prepared from celery stored under market conditions

and at 70F. Storage of celery for 2 weeks at 38 F resulted

in a significant increase in the peak area of limonene

(solvent extraction) while after 4 weeks' storage at 38F

the content of limonene (solvent extraction) was not differ-

ent from the control (Table 6). While no significant differ-

ence was observed between the limonene content (head-space)

of the control celery and samples stored 2 weeks at 38 F,

the treatment at 380F did result in a larger peak area and

all 4 C10H16 hydrocarbons analyzed followed the same general

trend.

The data substantiate the idea that storage of celery

at 70 F results in a high ratio of high/low boiling com-

ponents and presumably in a stronger (but not necessarily

more desirable) celery flavor, while storage at temperatures

between 38F and 50F result in lower ratios of high/low

boiling components and a possible reduction in strong celery

flavor. These assumptions should be confirmed through

rigorous experiments using organoleptic procedures.







Component Distribution

Considerable quantitative difference was observed

between the high and low boiling components in the chromato-

grams prepared from the various parts of the stalk. Table

14 presents the data for the ratio of high/low boiling com-

ponents. The mean ratio for the top was significantly

different from the middle and outer portions of the stalk,

while the ratio for the outer portion was significantly

higher than any other portion.


Table 14. Ratios of high/low boiling fractions from chro-
matograms of extracts prepared from various parts of the
celery plant.


Top Inner Bottom Middle Outer

0.66* 0.88 1.09 1.23 1.80




Those values not connected by a continuous line are
significantly different (0.05 level) according to
Duncan's multiple range test.


A qualitative difference was observed between the

chromatograms from the top (leafy) portion and all other

portions of the celery stalk. The chromatograms prepared

from the top portion had a large peak at 5 minutes, 25

seconds when chromatographed on a column of Apiezon L, and

5 minutes, 15 seconds when chromatographed on Carbowax 20 M.

This component occupied as much as 40 per cent of the total

peak area of these chromatograms and was absorbed by a







column containing boric acid which indicated it was an

alcohol (27). The odor of this compound was very ether-like

when detected at the exhaust port of the gas chromatograph.

Standard alcohols were chromatographed to aid in prediction

of carbon number and are plotted against retention times

in Figure 23.

The chromatogram from the top (leafy) extract is

presented in Figure 24. In addition to the unknown alcohol

previously mentioned, other qualitative variations in this

sample occurred near the retention time of 72 minutes. In

all other chromatograms the peak at 72 minutes appeared as

a major peak with a small shoulder. However, observation

of the chromatogram from the top (leafy) extract shows that

two peaks are completely resolved and present in large

quantities.

Chromatograms from celery extracts of inner and outer

portions are presented in Figure 25. While both fractions

showed a reduced amount of limonene as compared to other

fractions, the samples extracted from the inner portion of

the stalk had a larger quantity of the other components

with retention times of 10 to 40 minutes, yielding a lower

ratio of high/low boiling components. In addition to the

comparatively low amount of low boiling fraction in the

extracts of the outer portion, there was a larger amount

of those components with retention times of 60 to 90 minutes;

thus there was a higher ratio of high/low boiling components.

While few qualitative differences can be seen between these

















_- -Apiezon L

Carbowax 20 M


/I



)'^ ^ I ^. ~Unknown
+- I alcohol


5 6 7

Carbon number


Figure 23. Retention time of 5, 6, and 7 carbon
straight chain alcohols for polar and non-polar
chromatographic phases.






































S-i---!- -ri
71I





i__... ..
^Ij -I -\ T- !-- _
. --__,_! .- ,----! -















F +h~l: I !t -,
I- i _i c :II -i:-I :,
















T :tl -]K l ii i i :
,iiir: i






















c~w


4


Fl-4
0

g

43
0

0
P4



0

-4a
p.





*0





F4.
43
0







4-3
4'






5
A
41














09
k







.4.)



E
0
rd
i,
*
Kc

k


O


*o
0b
rl
C4)
p.

0"
nh











































































~-1
















II



B0


7'j


0




o















0








0
bo



0o






4-4
a
















0











0
54





00
(o
0i







a-


~g-c~ g






86

chromatograms, the fact that the inner portion of the stalk

resulted in a low, high/low ratio and the outer portion a

high, high/low ratio substantiates the differences in ratio

of high/low fractions previously mentioned.

Chromatograms from the middle and bottom portions of

the stalk are comparable to those of the storage treatment

extracts and little quantitative or qualitative difference

was observed. These results should be expected since the

celery used in the preparation of the storage treatment

extracts did not contain leafy portions and was primarily

composed of middle and bottom portions of the stalk.

While Pan (36, 37) and Hall (29) did not measure

differences in the volatile constituents of celery, their

research seems to substantiate these data. It was noted

that the strong bitter celery flavor was associated with the

outer petioles and with the more green portions of the stalk.

From these data, it is not possible to determine if the high

ratio of high/low boiling components is associated with a

strong celery flavor. However, according to the odor char-

acteristics presented in Table 3, this would appear true.

Also, in accordance with these data, Hall (21) found organ-

oleptic differences between middle and upper portions of

the outer and inner petioles of the same stalk.

Taste panel comparisons have been made between juice

expressed from the top (leafy) section of the celery plant

and that expressed from the bottom portion below the leaves

(14). The panel was able to differentiate between the two








samples at the 0.01 level of statistical significance.

However, these investigators indicated it was not possible

to differentiate between the two juices on the basis of the

gas chromatograms of their distillates.

Leaf alcohol (hex-cis-3 enol) is widely distributed

in green plants and is reported to be primarily responsible

for the odor of foliage plants (40, 42). Since the alcohol

in the celery extracts had a retention time similar to the

6 carbon group (Figure 23), it is possible that it is

closely related to the above-mentioned leaf alcohol and not

a specific product of celery. Noteworthy is the fact that

no peak corresponding to this retention time could be de-

tected when leafy portions of celery were analyzed using

head-space procedures.













SUMMARY AND CONCLUSIONS


Odor detection of the various components in the

celery extracts revealed that those components with reten-

tion times of 72 to 85 minutes possessed the characteristic

odor of celery. Analysis of the various chromatograms

revealed that two areas of high concentration were present

and were classified into high and low boiling components.

Further investigation of these areas revealed that net area

changes in the chromatograms could be detected by use of

the ratio of high/low boiling components.

Storage-induced flavor changes in celery stalks

appear to be quite temperature dependent when analyzed by

methods used in this research. Storage of celery at 70F,

whether fresh or previously stored at 38 F, resulted in a

decrease in the amount of C1IH16 hydrocarbons. Also,

complete flavor profile analysis showed a highly significant

decrease in total peak area of the chromatograms from

celery stored 5 days at 70F. While all components decreased

when celery was stored under these conditions, the ratio

of high to low boiling components increased significantly.

When celery was stored under market simulated con-

ditions (2 weeks at 45F and 1 subsequent week at 50F)

the ratio of high/low boiling components decreased

88







significantly. This decrease was probably due to increases

in the peak area of the C10H16 hydrocarbons. When celery

was stored at 380F or 450F, and analyzed using head-space

techniques, the C10H16 hydrocarbons content was higher

after 2 weeks' storage than for freshly harvested celery

(FWB); no further increase was noted with longer storage

at these temperatures. Storage at 50 F subsequent to

storage at 38 F also resulted in increases in these com-

ponents. Previous investigations (5) have established that

the basic mechanism of synthesis for these components is

the condensation of acetyl CoA and acetoacetyl CoA; however,

it is not known what limitation might cause a reversal in

the trend of component increase at 70 F storage.

Since other investigators have discussed the impor-

tance of the high boiling phthalides in celery flavor (10, 14,

15, 16, 20, 28, 32), the changes in the ratio of high/low

boiling fractions seem to have importance. In view of

these changes, it would appear that storage of celery at

70F would result in a stronger celery flavor, while storage

at 380F, 45 F, or combinations of these temperatures with 50F

storage would result in celery with less intense celery

flavor as compared to the same celery at harvest.

The same theory applies to the ratio of these com-

ponents in various portions of the stalk and is supported

by several investigators (14, 21, 22, 23). In general, the

strong portion of the celery flavor is associated with the

greener portions of the stalk, in particular the outer






90

petioles. This is further supported by the fact that the

inner (heart) petioles had a ratio of high/low boiling

fraction which was significantly lower than either the

middle or outer portions.

Further research should establish the organoleptic

association with these data. This would require rigorously

controlled conditions since celery stalks have considerable

changes in texture and appearance with long time storage.













BIBLIOGRAPHY


1. Allen, R. R. 1966. Selective removal of aldehydes
from complex mixtures during gas chromatography.
Anal. Chem. 38:1287.

2. Barton, D. H. R. and Juan X. DeVaries. 1963. The
constitution of sedanolide. Jour. of the Chem. Soc.
356:1916-1919.

3. Bellamy, L. J. 1959. The infrared spectra of complex
molecules. John Wiley and Sons, Inc. New York.

4. Berlingozzi, S. and L. Cione, as cited by Gold, H. J.
and C. W. Wilson, III. 1962. Alkylidene phthalides
and dihydrophthalides from celery. Jour. Org. Chem.
28:985.

5. Bernfeld, Peter. 1963. Biogenesis of natural com-
pounds. The Macmillan Company. New York.

6. Buttery, R. G. and R. Terabusgu. 1961. Gas liquid
chromatography of aroma of vegetables and fruits.
Direct injection of aqueous vapors. Anal. Chem.
33:1439-1441.
7. Ciamician, Giacomo and P. Silber. 1897. Uber die
hochsiedenden bestandtheile des sellerioles. Berichte
der Deutschen Chemischen Besellschaft. 30:492, 501,
1419, 1424, 1427.
8. Corbett, L. W. 1925. Physical and chemical changes
in celery during storage. Proc. Amer. Soc. Hort. Sci.
22:346-353.

9. Coyne, Dermot P. 1962. Chemical and physiological
changes in celery in relation to pithiness. Proc.
Amer. Soc. Hort. Sci. 81:341-346.

10. Crosby, D. G. and L. J. Anderson. 1963. The organic
constituents of food. II. Celery. Jour. of Food
Sci. 28:640-648.

11. Duncan, D. B. 1955. Multiple range and multiple F
tests. Biometrics 11:1-42.